67 files changed, 693 insertions, 530 deletions

diff --git a/src/common/fs/file.h b/src/common/fs/file.h
index 2c4ab4332..a4f7944cd 100644
--- a/src/common/fs/file.h
+++ b/src/common/fs/file.h
@@ -188,9 +188,8 @@ public:
 
 #ifdef _WIN32
     template <typename Path>
-    [[nodiscard]] void Open(const Path& path, FileAccessMode mode,
-                            FileType type = FileType::BinaryFile,
-                            FileShareFlag flag = FileShareFlag::ShareReadOnly) {
+    void Open(const Path& path, FileAccessMode mode, FileType type = FileType::BinaryFile,
+              FileShareFlag flag = FileShareFlag::ShareReadOnly) {
         using ValueType = typename Path::value_type;
         if constexpr (IsChar<ValueType>) {
             Open(ToU8String(path), mode, type, flag);
diff --git a/src/common/settings.h b/src/common/settings.h
index d01c0448c..9bee6e10f 100644
--- a/src/common/settings.h
+++ b/src/common/settings.h
@@ -554,6 +554,7 @@ struct Values {
     Setting<bool> use_docked_mode{true, "use_docked_mode"};
 
     BasicSetting<bool> enable_raw_input{false, "enable_raw_input"};
+    BasicSetting<bool> controller_navigation{true, "controller_navigation"};
 
     Setting<bool> vibration_enabled{true, "vibration_enabled"};
     Setting<bool> enable_accurate_vibrations{false, "enable_accurate_vibrations"};
diff --git a/src/common/wall_clock.cpp b/src/common/wall_clock.cpp
index 49830b8ab..9acf7551e 100644
--- a/src/common/wall_clock.cpp
+++ b/src/common/wall_clock.cpp
@@ -65,14 +65,20 @@ private:
 
 #ifdef ARCHITECTURE_x86_64
 
-std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
-                                                   u32 emulated_clock_frequency) {
+std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
+                                                   u64 emulated_clock_frequency) {
     const auto& caps = GetCPUCaps();
     u64 rtsc_frequency = 0;
     if (caps.invariant_tsc) {
         rtsc_frequency = EstimateRDTSCFrequency();
     }
-    if (rtsc_frequency == 0) {
+
+    // Fallback to StandardWallClock if the hardware TSC does not have the precision greater than:
+    // - A nanosecond
+    // - The emulated CPU frequency
+    // - The emulated clock counter frequency (CNTFRQ)
+    if (rtsc_frequency <= WallClock::NS_RATIO || rtsc_frequency <= emulated_cpu_frequency ||
+        rtsc_frequency <= emulated_clock_frequency) {
         return std::make_unique<StandardWallClock>(emulated_cpu_frequency,
                                                    emulated_clock_frequency);
     } else {
@@ -83,8 +89,8 @@ std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
 
 #else
 
-std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
-                                                   u32 emulated_clock_frequency) {
+std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
+                                                   u64 emulated_clock_frequency) {
     return std::make_unique<StandardWallClock>(emulated_cpu_frequency, emulated_clock_frequency);
 }
 
diff --git a/src/common/wall_clock.h b/src/common/wall_clock.h
index cef3e9499..874448c27 100644
--- a/src/common/wall_clock.h
+++ b/src/common/wall_clock.h
@@ -13,6 +13,10 @@ namespace Common {
 
 class WallClock {
 public:
+    static constexpr u64 NS_RATIO = 1'000'000'000;
+    static constexpr u64 US_RATIO = 1'000'000;
+    static constexpr u64 MS_RATIO = 1'000;
+
     virtual ~WallClock() = default;
 
     /// Returns current wall time in nanoseconds
@@ -49,7 +53,7 @@ private:
     bool is_native;
 };
 
-[[nodiscard]] std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
-                                                                 u32 emulated_clock_frequency);
+[[nodiscard]] std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
+                                                                 u64 emulated_clock_frequency);
 
 } // namespace Common
diff --git a/src/common/x64/native_clock.cpp b/src/common/x64/native_clock.cpp
index 82ee2c8a1..91b842829 100644
--- a/src/common/x64/native_clock.cpp
+++ b/src/common/x64/native_clock.cpp
@@ -47,9 +47,9 @@ NativeClock::NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequen
     _mm_mfence();
     time_point.inner.last_measure = __rdtsc();
     time_point.inner.accumulated_ticks = 0U;
-    ns_rtsc_factor = GetFixedPoint64Factor(1000000000, rtsc_frequency);
-    us_rtsc_factor = GetFixedPoint64Factor(1000000, rtsc_frequency);
-    ms_rtsc_factor = GetFixedPoint64Factor(1000, rtsc_frequency);
+    ns_rtsc_factor = GetFixedPoint64Factor(NS_RATIO, rtsc_frequency);
+    us_rtsc_factor = GetFixedPoint64Factor(US_RATIO, rtsc_frequency);
+    ms_rtsc_factor = GetFixedPoint64Factor(MS_RATIO, rtsc_frequency);
     clock_rtsc_factor = GetFixedPoint64Factor(emulated_clock_frequency, rtsc_frequency);
     cpu_rtsc_factor = GetFixedPoint64Factor(emulated_cpu_frequency, rtsc_frequency);
 }
diff --git a/src/common/x64/xbyak_abi.h b/src/common/x64/xbyak_abi.h
index 0ddf9b83e..87b3d63a4 100644
--- a/src/common/x64/xbyak_abi.h
+++ b/src/common/x64/xbyak_abi.h
@@ -37,12 +37,12 @@ constexpr Xbyak::Reg IndexToReg(size_t reg_index) {
     }
 }
 
-inline std::bitset<32> BuildRegSet(std::initializer_list<Xbyak::Reg> regs) {
-    std::bitset<32> bits;
+constexpr std::bitset<32> BuildRegSet(std::initializer_list<Xbyak::Reg> regs) {
+    size_t bits = 0;
     for (const Xbyak::Reg& reg : regs) {
-        bits[RegToIndex(reg)] = true;
+        bits |= size_t{1} << RegToIndex(reg);
     }
-    return bits;
+    return {bits};
 }
 
 constexpr inline std::bitset<32> ABI_ALL_GPRS(0x0000FFFF);
@@ -57,7 +57,7 @@ constexpr inline Xbyak::Reg ABI_PARAM2 = Xbyak::util::rdx;
 constexpr inline Xbyak::Reg ABI_PARAM3 = Xbyak::util::r8;
 constexpr inline Xbyak::Reg ABI_PARAM4 = Xbyak::util::r9;
 
-const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rcx,
     Xbyak::util::rdx,
@@ -74,7 +74,7 @@ const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     Xbyak::util::xmm5,
 });
 
-const std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rbx,
     Xbyak::util::rsi,
@@ -108,7 +108,7 @@ constexpr inline Xbyak::Reg ABI_PARAM2 = Xbyak::util::rsi;
 constexpr inline Xbyak::Reg ABI_PARAM3 = Xbyak::util::rdx;
 constexpr inline Xbyak::Reg ABI_PARAM4 = Xbyak::util::rcx;
 
-const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rcx,
     Xbyak::util::rdx,
@@ -137,7 +137,7 @@ const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     Xbyak::util::xmm15,
 });
 
-const std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rbx,
     Xbyak::util::rbp,
diff --git a/src/core/hid/emulated_controller.cpp b/src/core/hid/emulated_controller.cpp
index d12037b11..a7cdf45e6 100644
--- a/src/core/hid/emulated_controller.cpp
+++ b/src/core/hid/emulated_controller.cpp
@@ -749,6 +749,7 @@ void EmulatedController::SetMotion(const Common::Input::CallbackStatus& callback
         raw_status.gyro.y.value,
         raw_status.gyro.z.value,
     });
+    emulated.SetGyroThreshold(raw_status.gyro.x.properties.threshold);
     emulated.UpdateRotation(raw_status.delta_timestamp);
     emulated.UpdateOrientation(raw_status.delta_timestamp);
     force_update_motion = raw_status.force_update;
diff --git a/src/core/hid/motion_input.cpp b/src/core/hid/motion_input.cpp
index 6e126be19..05042fd99 100644
--- a/src/core/hid/motion_input.cpp
+++ b/src/core/hid/motion_input.cpp
@@ -10,7 +10,7 @@ namespace Core::HID {
 MotionInput::MotionInput() {
     // Initialize PID constants with default values
     SetPID(0.3f, 0.005f, 0.0f);
-    SetGyroThreshold(0.00005f);
+    SetGyroThreshold(0.007f);
 }
 
 void MotionInput::SetPID(f32 new_kp, f32 new_ki, f32 new_kd) {
@@ -31,7 +31,7 @@ void MotionInput::SetGyroscope(const Common::Vec3f& gyroscope) {
         gyro_bias = (gyro_bias * 0.9999f) + (gyroscope * 0.0001f);
     }
 
-    if (gyro.Length2() < gyro_threshold) {
+    if (gyro.Length() < gyro_threshold) {
         gyro = {};
     } else {
         only_accelerometer = false;
diff --git a/src/core/hle/kernel/k_affinity_mask.h b/src/core/hle/kernel/k_affinity_mask.h
index b906895fc..cf704ce87 100644
--- a/src/core/hle/kernel/k_affinity_mask.h
+++ b/src/core/hle/kernel/k_affinity_mask.h
@@ -31,8 +31,6 @@ public:
     }
 
     constexpr void SetAffinity(s32 core, bool set) {
-        ASSERT(0 <= core && core < static_cast<s32>(Core::Hardware::NUM_CPU_CORES));
-
         if (set) {
             this->mask |= GetCoreBit(core);
         } else {
diff --git a/src/core/hle/kernel/k_page_table.cpp b/src/core/hle/kernel/k_page_table.cpp
index b650ea31d..2ebbc0819 100644
--- a/src/core/hle/kernel/k_page_table.cpp
+++ b/src/core/hle/kernel/k_page_table.cpp
@@ -276,22 +276,23 @@ ResultCode KPageTable::InitializeForProcess(FileSys::ProgramAddressSpaceType as_
 
 ResultCode KPageTable::MapProcessCode(VAddr addr, std::size_t num_pages, KMemoryState state,
                                       KMemoryPermission perm) {
-    std::lock_guard lock{page_table_lock};
-
     const u64 size{num_pages * PageSize};
 
-    if (!CanContain(addr, size, state)) {
-        return ResultInvalidCurrentMemory;
-    }
+    // Validate the mapping request.
+    R_UNLESS(this->CanContain(addr, size, state), ResultInvalidCurrentMemory);
 
-    if (IsRegionMapped(addr, size)) {
-        return ResultInvalidCurrentMemory;
-    }
+    // Lock the table.
+    std::lock_guard lock{page_table_lock};
+
+    // Verify that the destination memory is unmapped.
+    R_TRY(this->CheckMemoryState(addr, size, KMemoryState::All, KMemoryState::Free,
+                                 KMemoryPermission::None, KMemoryPermission::None,
+                                 KMemoryAttribute::None, KMemoryAttribute::None));
 
     KPageLinkedList page_linked_list;
-    CASCADE_CODE(system.Kernel().MemoryManager().Allocate(page_linked_list, num_pages, memory_pool,
-                                                          allocation_option));
-    CASCADE_CODE(Operate(addr, num_pages, page_linked_list, OperationType::MapGroup));
+    R_TRY(system.Kernel().MemoryManager().Allocate(page_linked_list, num_pages, memory_pool,
+                                                   allocation_option));
+    R_TRY(Operate(addr, num_pages, page_linked_list, OperationType::MapGroup));
 
     block_manager->Update(addr, num_pages, state, perm);
 
@@ -395,39 +396,12 @@ ResultCode KPageTable::UnmapProcessMemory(VAddr dst_addr, std::size_t size,
 
     return ResultSuccess;
 }
-void KPageTable::MapPhysicalMemory(KPageLinkedList& page_linked_list, VAddr start, VAddr end) {
-    auto node{page_linked_list.Nodes().begin()};
-    PAddr map_addr{node->GetAddress()};
-    std::size_t src_num_pages{node->GetNumPages()};
-
-    block_manager->IterateForRange(start, end, [&](const KMemoryInfo& info) {
-        if (info.state != KMemoryState::Free) {
-            return;
-        }
-
-        std::size_t dst_num_pages{GetSizeInRange(info, start, end) / PageSize};
-        VAddr dst_addr{GetAddressInRange(info, start)};
-
-        while (dst_num_pages) {
-            if (!src_num_pages) {
-                node = std::next(node);
-                map_addr = node->GetAddress();
-                src_num_pages = node->GetNumPages();
-            }
-
-            const std::size_t num_pages{std::min(src_num_pages, dst_num_pages)};
-            Operate(dst_addr, num_pages, KMemoryPermission::UserReadWrite, OperationType::Map,
-                    map_addr);
-
-            dst_addr += num_pages * PageSize;
-            map_addr += num_pages * PageSize;
-            src_num_pages -= num_pages;
-            dst_num_pages -= num_pages;
-        }
-    });
-}
 
 ResultCode KPageTable::MapPhysicalMemory(VAddr addr, std::size_t size) {
+    // Lock the physical memory lock.
+    std::lock_guard phys_lk(map_physical_memory_lock);
+
+    // Lock the table.
     std::lock_guard lock{page_table_lock};
 
     std::size_t mapped_size{};
@@ -463,7 +437,35 @@ ResultCode KPageTable::MapPhysicalMemory(VAddr addr, std::size_t size) {
     // We succeeded, so commit the memory reservation.
     memory_reservation.Commit();
 
-    MapPhysicalMemory(page_linked_list, addr, end_addr);
+    // Map the memory.
+    auto node{page_linked_list.Nodes().begin()};
+    PAddr map_addr{node->GetAddress()};
+    std::size_t src_num_pages{node->GetNumPages()};
+    block_manager->IterateForRange(addr, end_addr, [&](const KMemoryInfo& info) {
+        if (info.state != KMemoryState::Free) {
+            return;
+        }
+
+        std::size_t dst_num_pages{GetSizeInRange(info, addr, end_addr) / PageSize};
+        VAddr dst_addr{GetAddressInRange(info, addr)};
+
+        while (dst_num_pages) {
+            if (!src_num_pages) {
+                node = std::next(node);
+                map_addr = node->GetAddress();
+                src_num_pages = node->GetNumPages();
+            }
+
+            const std::size_t num_pages{std::min(src_num_pages, dst_num_pages)};
+            Operate(dst_addr, num_pages, KMemoryPermission::UserReadWrite, OperationType::Map,
+                    map_addr);
+
+            dst_addr += num_pages * PageSize;
+            map_addr += num_pages * PageSize;
+            src_num_pages -= num_pages;
+            dst_num_pages -= num_pages;
+        }
+    });
 
     mapped_physical_memory_size += remaining_size;
 
@@ -503,23 +505,8 @@ ResultCode KPageTable::UnmapPhysicalMemory(VAddr addr, std::size_t size) {
         return ResultSuccess;
     }
 
-    CASCADE_CODE(UnmapMemory(addr, size));
-
-    auto process{system.Kernel().CurrentProcess()};
-    process->GetResourceLimit()->Release(LimitableResource::PhysicalMemory, mapped_size);
-    mapped_physical_memory_size -= mapped_size;
-
-    return ResultSuccess;
-}
-
-ResultCode KPageTable::UnmapMemory(VAddr addr, std::size_t size) {
-    std::lock_guard lock{page_table_lock};
-
-    const VAddr end_addr{addr + size};
-    ResultCode result{ResultSuccess};
-    KPageLinkedList page_linked_list;
-
     // Unmap each region within the range
+    KPageLinkedList page_linked_list;
     block_manager->IterateForRange(addr, end_addr, [&](const KMemoryInfo& info) {
         if (info.state == KMemoryState::Normal) {
             const std::size_t block_size{GetSizeInRange(info, addr, end_addr)};
@@ -535,7 +522,6 @@ ResultCode KPageTable::UnmapMemory(VAddr addr, std::size_t size) {
             }
         }
     });
-
     if (result.IsError()) {
         return result;
     }
@@ -546,10 +532,14 @@ ResultCode KPageTable::UnmapMemory(VAddr addr, std::size_t size) {
 
     block_manager->Update(addr, num_pages, KMemoryState::Free);
 
+    auto process{system.Kernel().CurrentProcess()};
+    process->GetResourceLimit()->Release(LimitableResource::PhysicalMemory, mapped_size);
+    mapped_physical_memory_size -= mapped_size;
+
     return ResultSuccess;
 }
 
-ResultCode KPageTable::Map(VAddr dst_addr, VAddr src_addr, std::size_t size) {
+ResultCode KPageTable::MapMemory(VAddr dst_addr, VAddr src_addr, std::size_t size) {
     std::lock_guard lock{page_table_lock};
 
     KMemoryState src_state{};
@@ -588,7 +578,7 @@ ResultCode KPageTable::Map(VAddr dst_addr, VAddr src_addr, std::size_t size) {
     return ResultSuccess;
 }
 
-ResultCode KPageTable::Unmap(VAddr dst_addr, VAddr src_addr, std::size_t size) {
+ResultCode KPageTable::UnmapMemory(VAddr dst_addr, VAddr src_addr, std::size_t size) {
     std::lock_guard lock{page_table_lock};
 
     KMemoryState src_state{};
@@ -652,24 +642,26 @@ ResultCode KPageTable::MapPages(VAddr addr, const KPageLinkedList& page_linked_l
     return ResultSuccess;
 }
 
-ResultCode KPageTable::MapPages(VAddr addr, KPageLinkedList& page_linked_list, KMemoryState state,
-                                KMemoryPermission perm) {
-    std::lock_guard lock{page_table_lock};
-
+ResultCode KPageTable::MapPages(VAddr address, KPageLinkedList& page_linked_list,
+                                KMemoryState state, KMemoryPermission perm) {
+    // Check that the map is in range.
     const std::size_t num_pages{page_linked_list.GetNumPages()};
     const std::size_t size{num_pages * PageSize};
+    R_UNLESS(this->CanContain(address, size, state), ResultInvalidCurrentMemory);
 
-    if (!CanContain(addr, size, state)) {
-        return ResultInvalidCurrentMemory;
-    }
+    // Lock the table.
+    std::lock_guard lock{page_table_lock};
 
-    if (IsRegionMapped(addr, num_pages * PageSize)) {
-        return ResultInvalidCurrentMemory;
-    }
+    // Check the memory state.
+    R_TRY(this->CheckMemoryState(address, size, KMemoryState::All, KMemoryState::Free,
+                                 KMemoryPermission::None, KMemoryPermission::None,
+                                 KMemoryAttribute::None, KMemoryAttribute::None));
 
-    CASCADE_CODE(MapPages(addr, page_linked_list, perm));
+    // Map the pages.
+    R_TRY(MapPages(address, page_linked_list, perm));
 
-    block_manager->Update(addr, num_pages, state, perm);
+    // Update the blocks.
+    block_manager->Update(address, num_pages, state, perm);
 
     return ResultSuccess;
 }
@@ -693,21 +685,23 @@ ResultCode KPageTable::UnmapPages(VAddr addr, const KPageLinkedList& page_linked
 
 ResultCode KPageTable::UnmapPages(VAddr addr, KPageLinkedList& page_linked_list,
                                   KMemoryState state) {
-    std::lock_guard lock{page_table_lock};
-
+    // Check that the unmap is in range.
     const std::size_t num_pages{page_linked_list.GetNumPages()};
     const std::size_t size{num_pages * PageSize};
+    R_UNLESS(this->Contains(addr, size), ResultInvalidCurrentMemory);
 
-    if (!CanContain(addr, size, state)) {
-        return ResultInvalidCurrentMemory;
-    }
+    // Lock the table.
+    std::lock_guard lock{page_table_lock};
 
-    if (IsRegionMapped(addr, num_pages * PageSize)) {
-        return ResultInvalidCurrentMemory;
-    }
+    // Check the memory state.
+    R_TRY(this->CheckMemoryState(addr, size, KMemoryState::All, state, KMemoryPermission::None,
+                                 KMemoryPermission::None, KMemoryAttribute::All,
+                                 KMemoryAttribute::None));
 
-    CASCADE_CODE(UnmapPages(addr, page_linked_list));
+    // Perform the unmap.
+    R_TRY(UnmapPages(addr, page_linked_list));
 
+    // Update the blocks.
     block_manager->Update(addr, num_pages, state, KMemoryPermission::None);
 
     return ResultSuccess;
@@ -765,7 +759,6 @@ ResultCode KPageTable::SetProcessMemoryPermission(VAddr addr, std::size_t size,
 
     // Ensure cache coherency, if we're setting pages as executable.
     if (is_x) {
-        // Memory execution state is changing, invalidate CPU cache range
         system.InvalidateCpuInstructionCacheRange(addr, size);
     }
 
@@ -793,12 +786,12 @@ ResultCode KPageTable::ReserveTransferMemory(VAddr addr, std::size_t size, KMemo
     KMemoryState state{};
     KMemoryAttribute attribute{};
 
-    CASCADE_CODE(CheckMemoryState(
-        &state, nullptr, &attribute, nullptr, addr, size,
-        KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
-        KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted, KMemoryPermission::All,
-        KMemoryPermission::UserReadWrite, KMemoryAttribute::Mask, KMemoryAttribute::None,
-        KMemoryAttribute::IpcAndDeviceMapped));
+    R_TRY(CheckMemoryState(&state, nullptr, &attribute, nullptr, addr, size,
+                           KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
+                           KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
+                           KMemoryPermission::All, KMemoryPermission::UserReadWrite,
+                           KMemoryAttribute::Mask, KMemoryAttribute::None,
+                           KMemoryAttribute::IpcAndDeviceMapped));
 
     block_manager->Update(addr, size / PageSize, state, perm, attribute | KMemoryAttribute::Locked);
 
@@ -810,12 +803,11 @@ ResultCode KPageTable::ResetTransferMemory(VAddr addr, std::size_t size) {
 
     KMemoryState state{};
 
-    CASCADE_CODE(
-        CheckMemoryState(&state, nullptr, nullptr, nullptr, addr, size,
-                         KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
-                         KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
-                         KMemoryPermission::None, KMemoryPermission::None, KMemoryAttribute::Mask,
-                         KMemoryAttribute::Locked, KMemoryAttribute::IpcAndDeviceMapped));
+    R_TRY(CheckMemoryState(&state, nullptr, nullptr, nullptr, addr, size,
+                           KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
+                           KMemoryState::FlagCanTransfer | KMemoryState::FlagReferenceCounted,
+                           KMemoryPermission::None, KMemoryPermission::None, KMemoryAttribute::Mask,
+                           KMemoryAttribute::Locked, KMemoryAttribute::IpcAndDeviceMapped));
 
     block_manager->Update(addr, size / PageSize, state, KMemoryPermission::UserReadWrite);
     return ResultSuccess;
@@ -871,8 +863,9 @@ ResultCode KPageTable::SetMemoryAttribute(VAddr addr, std::size_t size, u32 mask
         AttributeTestMask, KMemoryAttribute::None, ~AttributeTestMask));
 
     // Determine the new attribute.
-    const auto new_attr = ((old_attr & static_cast<KMemoryAttribute>(~mask)) |
-                           static_cast<KMemoryAttribute>(attr & mask));
+    const KMemoryAttribute new_attr =
+        static_cast<KMemoryAttribute>(((old_attr & static_cast<KMemoryAttribute>(~mask)) |
+                                       static_cast<KMemoryAttribute>(attr & mask)));
 
     // Perform operation.
     this->Operate(addr, num_pages, old_perm, OperationType::ChangePermissionsAndRefresh);
@@ -896,6 +889,9 @@ ResultCode KPageTable::SetMaxHeapSize(std::size_t size) {
 }
 
 ResultCode KPageTable::SetHeapSize(VAddr* out, std::size_t size) {
+    // Lock the physical memory lock.
+    std::lock_guard phys_lk(map_physical_memory_lock);
+
     // Try to perform a reduction in heap, instead of an extension.
     VAddr cur_address{};
     std::size_t allocation_size{};
@@ -1025,12 +1021,12 @@ ResultVal<VAddr> KPageTable::AllocateAndMapMemory(std::size_t needed_num_pages,
     }
 
     if (is_map_only) {
-        CASCADE_CODE(Operate(addr, needed_num_pages, perm, OperationType::Map, map_addr));
+        R_TRY(Operate(addr, needed_num_pages, perm, OperationType::Map, map_addr));
     } else {
         KPageLinkedList page_group;
-        CASCADE_CODE(system.Kernel().MemoryManager().Allocate(page_group, needed_num_pages,
-                                                              memory_pool, allocation_option));
-        CASCADE_CODE(Operate(addr, needed_num_pages, page_group, OperationType::MapGroup));
+        R_TRY(system.Kernel().MemoryManager().Allocate(page_group, needed_num_pages, memory_pool,
+                                                       allocation_option));
+        R_TRY(Operate(addr, needed_num_pages, page_group, OperationType::MapGroup));
     }
 
     block_manager->Update(addr, needed_num_pages, state, perm);
@@ -1186,7 +1182,7 @@ VAddr KPageTable::AllocateVirtualMemory(VAddr start, std::size_t region_num_page
 
 ResultCode KPageTable::Operate(VAddr addr, std::size_t num_pages, const KPageLinkedList& page_group,
                                OperationType operation) {
-    std::lock_guard lock{page_table_lock};
+    ASSERT(this->IsLockedByCurrentThread());
 
     ASSERT(Common::IsAligned(addr, PageSize));
     ASSERT(num_pages > 0);
@@ -1211,7 +1207,7 @@ ResultCode KPageTable::Operate(VAddr addr, std::size_t num_pages, const KPageLin
 
 ResultCode KPageTable::Operate(VAddr addr, std::size_t num_pages, KMemoryPermission perm,
                                OperationType operation, PAddr map_addr) {
-    std::lock_guard lock{page_table_lock};
+    ASSERT(this->IsLockedByCurrentThread());
 
     ASSERT(num_pages > 0);
     ASSERT(Common::IsAligned(addr, PageSize));
diff --git a/src/core/hle/kernel/k_page_table.h b/src/core/hle/kernel/k_page_table.h
index f67986e91..60ae9b9e8 100644
--- a/src/core/hle/kernel/k_page_table.h
+++ b/src/core/hle/kernel/k_page_table.h
@@ -37,9 +37,8 @@ public:
                                   VAddr src_addr);
     ResultCode MapPhysicalMemory(VAddr addr, std::size_t size);
     ResultCode UnmapPhysicalMemory(VAddr addr, std::size_t size);
-    ResultCode UnmapMemory(VAddr addr, std::size_t size);
-    ResultCode Map(VAddr dst_addr, VAddr src_addr, std::size_t size);
-    ResultCode Unmap(VAddr dst_addr, VAddr src_addr, std::size_t size);
+    ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
+    ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
     ResultCode MapPages(VAddr addr, KPageLinkedList& page_linked_list, KMemoryState state,
                         KMemoryPermission perm);
     ResultCode UnmapPages(VAddr addr, KPageLinkedList& page_linked_list, KMemoryState state);
@@ -88,7 +87,6 @@ private:
     ResultCode MapPages(VAddr addr, const KPageLinkedList& page_linked_list,
                         KMemoryPermission perm);
     ResultCode UnmapPages(VAddr addr, const KPageLinkedList& page_linked_list);
-    void MapPhysicalMemory(KPageLinkedList& page_linked_list, VAddr start, VAddr end);
     bool IsRegionMapped(VAddr address, u64 size);
     bool IsRegionContiguous(VAddr addr, u64 size) const;
     void AddRegionToPages(VAddr start, std::size_t num_pages, KPageLinkedList& page_linked_list);
@@ -148,6 +146,7 @@ private:
     }
 
     std::recursive_mutex page_table_lock;
+    std::mutex map_physical_memory_lock;
     std::unique_ptr<KMemoryBlockManager> block_manager;
 
 public:
@@ -249,7 +248,9 @@ public:
         return !IsOutsideASLRRegion(address, size);
     }
     constexpr PAddr GetPhysicalAddr(VAddr addr) {
-        return page_table_impl.backing_addr[addr >> PageBits] + addr;
+        const auto backing_addr = page_table_impl.backing_addr[addr >> PageBits];
+        ASSERT(backing_addr);
+        return backing_addr + addr;
     }
     constexpr bool Contains(VAddr addr) const {
         return address_space_start <= addr && addr <= address_space_end - 1;
diff --git a/src/core/hle/kernel/k_priority_queue.h b/src/core/hle/kernel/k_priority_queue.h
index 0b894c8cf..bd779739d 100644
--- a/src/core/hle/kernel/k_priority_queue.h
+++ b/src/core/hle/kernel/k_priority_queue.h
@@ -258,7 +258,7 @@ private:
 
 private:
     constexpr void ClearAffinityBit(u64& affinity, s32 core) {
-        affinity &= ~(u64(1) << core);
+        affinity &= ~(UINT64_C(1) << core);
     }
 
     constexpr s32 GetNextCore(u64& affinity) {
diff --git a/src/core/hle/kernel/k_process.cpp b/src/core/hle/kernel/k_process.cpp
index 265ac6fa1..85c506979 100644
--- a/src/core/hle/kernel/k_process.cpp
+++ b/src/core/hle/kernel/k_process.cpp
@@ -146,6 +146,13 @@ ResultCode KProcess::Initialize(KProcess* process, Core::System& system, std::st
     // Open a reference to the resource limit.
     process->resource_limit->Open();
 
+    // Clear remaining fields.
+    process->num_running_threads = 0;
+    process->is_signaled = false;
+    process->exception_thread = nullptr;
+    process->is_suspended = false;
+    process->schedule_count = 0;
+
     return ResultSuccess;
 }
 
@@ -157,20 +164,17 @@ KResourceLimit* KProcess::GetResourceLimit() const {
     return resource_limit;
 }
 
-void KProcess::IncrementThreadCount() {
-    ASSERT(num_threads >= 0);
-    num_created_threads++;
-
-    if (const auto count = ++num_threads; count > peak_num_threads) {
-        peak_num_threads = count;
-    }
+void KProcess::IncrementRunningThreadCount() {
+    ASSERT(num_running_threads.load() >= 0);
+    ++num_running_threads;
 }
 
-void KProcess::DecrementThreadCount() {
-    ASSERT(num_threads > 0);
+void KProcess::DecrementRunningThreadCount() {
+    ASSERT(num_running_threads.load() > 0);
 
-    if (const auto count = --num_threads; count == 0) {
-        LOG_WARNING(Kernel, "Process termination is not fully implemented.");
+    if (const auto prev = num_running_threads--; prev == 1) {
+        // TODO(bunnei): Process termination to be implemented when multiprocess is supported.
+        UNIMPLEMENTED_MSG("KProcess termination is not implemennted!");
     }
 }
 
diff --git a/src/core/hle/kernel/k_process.h b/src/core/hle/kernel/k_process.h
index c2a672021..38b446350 100644
--- a/src/core/hle/kernel/k_process.h
+++ b/src/core/hle/kernel/k_process.h
@@ -235,8 +235,8 @@ public:
         ++schedule_count;
     }
 
-    void IncrementThreadCount();
-    void DecrementThreadCount();
+    void IncrementRunningThreadCount();
+    void DecrementRunningThreadCount();
 
     void SetRunningThread(s32 core, KThread* thread, u64 idle_count) {
         running_threads[core] = thread;
@@ -473,9 +473,7 @@ private:
     bool is_suspended{};
     bool is_initialized{};
 
-    std::atomic<s32> num_created_threads{};
-    std::atomic<u16> num_threads{};
-    u16 peak_num_threads{};
+    std::atomic<u16> num_running_threads{};
 
     std::array<KThread*, Core::Hardware::NUM_CPU_CORES> running_threads{};
     std::array<u64, Core::Hardware::NUM_CPU_CORES> running_thread_idle_counts{};
diff --git a/src/core/hle/kernel/k_scheduler.cpp b/src/core/hle/kernel/k_scheduler.cpp
index b32d4f285..c96520828 100644
--- a/src/core/hle/kernel/k_scheduler.cpp
+++ b/src/core/hle/kernel/k_scheduler.cpp
@@ -710,23 +710,19 @@ void KScheduler::Unload(KThread* thread) {
 }
 
 void KScheduler::Reload(KThread* thread) {
-    LOG_TRACE(Kernel, "core {}, reload thread {}", core_id, thread ? thread->GetName() : "nullptr");
+    LOG_TRACE(Kernel, "core {}, reload thread {}", core_id, thread->GetName());
 
-    if (thread) {
-        ASSERT_MSG(thread->GetState() == ThreadState::Runnable, "Thread must be runnable.");
-
-        Core::ARM_Interface& cpu_core = system.ArmInterface(core_id);
-        cpu_core.LoadContext(thread->GetContext32());
-        cpu_core.LoadContext(thread->GetContext64());
-        cpu_core.SetTlsAddress(thread->GetTLSAddress());
-        cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0());
-        cpu_core.ClearExclusiveState();
-    }
+    Core::ARM_Interface& cpu_core = system.ArmInterface(core_id);
+    cpu_core.LoadContext(thread->GetContext32());
+    cpu_core.LoadContext(thread->GetContext64());
+    cpu_core.SetTlsAddress(thread->GetTLSAddress());
+    cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0());
+    cpu_core.ClearExclusiveState();
 }
 
 void KScheduler::SwitchContextStep2() {
     // Load context of new thread
-    Reload(current_thread.load());
+    Reload(GetCurrentThread());
 
     RescheduleCurrentCore();
 }
@@ -735,13 +731,17 @@ void KScheduler::ScheduleImpl() {
     KThread* previous_thread = GetCurrentThread();
     KThread* next_thread = state.highest_priority_thread;
 
-    state.needs_scheduling = false;
+    state.needs_scheduling.store(false);
 
     // We never want to schedule a null thread, so use the idle thread if we don't have a next.
     if (next_thread == nullptr) {
         next_thread = idle_thread;
     }
 
+    if (next_thread->GetCurrentCore() != core_id) {
+        next_thread->SetCurrentCore(core_id);
+    }
+
     // We never want to schedule a dummy thread, as these are only used by host threads for locking.
     if (next_thread->GetThreadType() == ThreadType::Dummy) {
         ASSERT_MSG(false, "Dummy threads should never be scheduled!");
@@ -755,14 +755,8 @@ void KScheduler::ScheduleImpl() {
         return;
     }
 
-    if (next_thread->GetCurrentCore() != core_id) {
-        next_thread->SetCurrentCore(core_id);
-    }
-
-    current_thread.store(next_thread);
-
+    // Update the CPU time tracking variables.
     KProcess* const previous_process = system.Kernel().CurrentProcess();
-
     UpdateLastContextSwitchTime(previous_thread, previous_process);
 
     // Save context for previous thread
@@ -770,6 +764,10 @@ void KScheduler::ScheduleImpl() {
 
     std::shared_ptr<Common::Fiber>* old_context;
     old_context = &previous_thread->GetHostContext();
+
+    // Set the new thread.
+    current_thread.store(next_thread);
+
     guard.Unlock();
 
     Common::Fiber::YieldTo(*old_context, *switch_fiber);
@@ -797,8 +795,8 @@ void KScheduler::SwitchToCurrent() {
         do {
             auto next_thread = current_thread.load();
             if (next_thread != nullptr) {
-                next_thread->context_guard.Lock();
-                if (next_thread->GetRawState() != ThreadState::Runnable) {
+                const auto locked = next_thread->context_guard.TryLock();
+                if (state.needs_scheduling.load()) {
                     next_thread->context_guard.Unlock();
                     break;
                 }
@@ -806,6 +804,9 @@ void KScheduler::SwitchToCurrent() {
                     next_thread->context_guard.Unlock();
                     break;
                 }
+                if (!locked) {
+                    continue;
+                }
             }
             auto thread = next_thread ? next_thread : idle_thread;
             Common::Fiber::YieldTo(switch_fiber, *thread->GetHostContext());
diff --git a/src/core/hle/kernel/k_thread.cpp b/src/core/hle/kernel/k_thread.cpp
index f42abb8a1..de3ffe0c7 100644
--- a/src/core/hle/kernel/k_thread.cpp
+++ b/src/core/hle/kernel/k_thread.cpp
@@ -215,7 +215,6 @@ ResultCode KThread::Initialize(KThreadFunction func, uintptr_t arg, VAddr user_s
 
         parent = owner;
         parent->Open();
-        parent->IncrementThreadCount();
     }
 
     // Initialize thread context.
@@ -327,11 +326,6 @@ void KThread::Finalize() {
         }
     }
 
-    // Decrement the parent process's thread count.
-    if (parent != nullptr) {
-        parent->DecrementThreadCount();
-    }
-
     // Perform inherited finalization.
     KSynchronizationObject::Finalize();
 }
@@ -1011,7 +1005,7 @@ ResultCode KThread::Run() {
             if (IsUserThread() && IsSuspended()) {
                 this->UpdateState();
             }
-            owner->IncrementThreadCount();
+            owner->IncrementRunningThreadCount();
         }
 
         // Set our state and finish.
@@ -1026,10 +1020,11 @@ ResultCode KThread::Run() {
 void KThread::Exit() {
     ASSERT(this == GetCurrentThreadPointer(kernel));
 
-    // Release the thread resource hint from parent.
+    // Release the thread resource hint, running thread count from parent.
     if (parent != nullptr) {
         parent->GetResourceLimit()->Release(Kernel::LimitableResource::Threads, 0, 1);
         resource_limit_release_hint = true;
+        parent->DecrementRunningThreadCount();
     }
 
     // Perform termination.
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index c7f5140f4..4f7aebf3f 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -230,7 +230,7 @@ static ResultCode MapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr
         return result;
     }
 
-    return page_table.Map(dst_addr, src_addr, size);
+    return page_table.MapMemory(dst_addr, src_addr, size);
 }
 
 static ResultCode MapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) {
@@ -249,7 +249,7 @@ static ResultCode UnmapMemory(Core::System& system, VAddr dst_addr, VAddr src_ad
         return result;
     }
 
-    return page_table.Unmap(dst_addr, src_addr, size);
+    return page_table.UnmapMemory(dst_addr, src_addr, size);
 }
 
 static ResultCode UnmapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) {
@@ -2613,7 +2613,7 @@ static const FunctionDef SVC_Table_32[] = {
     {0x33, SvcWrap32<GetThreadContext32>, "GetThreadContext32"},
     {0x34, SvcWrap32<WaitForAddress32>, "WaitForAddress32"},
     {0x35, SvcWrap32<SignalToAddress32>, "SignalToAddress32"},
-    {0x36, nullptr, "Unknown"},
+    {0x36, SvcWrap32<SynchronizePreemptionState>, "SynchronizePreemptionState32"},
     {0x37, nullptr, "Unknown"},
     {0x38, nullptr, "Unknown"},
     {0x39, nullptr, "Unknown"},
diff --git a/src/input_common/drivers/sdl_driver.cpp b/src/input_common/drivers/sdl_driver.cpp
index ed6281772..577bf5c31 100644
--- a/src/input_common/drivers/sdl_driver.cpp
+++ b/src/input_common/drivers/sdl_driver.cpp
@@ -109,8 +109,9 @@ public:
 
     bool HasHDRumble() const {
         if (sdl_controller) {
-            return (SDL_GameControllerGetType(sdl_controller.get()) ==
-                    SDL_CONTROLLER_TYPE_NINTENDO_SWITCH_PRO);
+            const auto type = SDL_GameControllerGetType(sdl_controller.get());
+            return (type == SDL_CONTROLLER_TYPE_NINTENDO_SWITCH_PRO) ||
+                   (type == SDL_CONTROLLER_TYPE_PS5);
         }
         return false;
     }
diff --git a/src/input_common/drivers/udp_client.cpp b/src/input_common/drivers/udp_client.cpp
index 9aaeb91be..d1cdb1ab2 100644
--- a/src/input_common/drivers/udp_client.cpp
+++ b/src/input_common/drivers/udp_client.cpp
@@ -339,7 +339,7 @@ void UDPClient::StartCommunication(std::size_t client, const std::string& host,
     }
 }
 
-const PadIdentifier UDPClient::GetPadIdentifier(std::size_t pad_index) const {
+PadIdentifier UDPClient::GetPadIdentifier(std::size_t pad_index) const {
     const std::size_t client = pad_index / PADS_PER_CLIENT;
     return {
         .guid = clients[client].uuid,
@@ -348,9 +348,9 @@ const PadIdentifier UDPClient::GetPadIdentifier(std::size_t pad_index) const {
     };
 }
 
-const Common::UUID UDPClient::GetHostUUID(const std::string host) const {
-    const auto ip = boost::asio::ip::address_v4::from_string(host);
-    const auto hex_host = fmt::format("{:06x}", ip.to_ulong());
+Common::UUID UDPClient::GetHostUUID(const std::string& host) const {
+    const auto ip = boost::asio::ip::make_address_v4(host);
+    const auto hex_host = fmt::format("{:06x}", ip.to_uint());
     return Common::UUID{hex_host};
 }
 
diff --git a/src/input_common/drivers/udp_client.h b/src/input_common/drivers/udp_client.h
index 61a1fff37..30d7c2682 100644
--- a/src/input_common/drivers/udp_client.h
+++ b/src/input_common/drivers/udp_client.h
@@ -145,8 +145,8 @@ private:
     void OnPortInfo(Response::PortInfo);
     void OnPadData(Response::PadData, std::size_t client);
     void StartCommunication(std::size_t client, const std::string& host, u16 port);
-    const PadIdentifier GetPadIdentifier(std::size_t pad_index) const;
-    const Common::UUID GetHostUUID(const std::string host) const;
+    PadIdentifier GetPadIdentifier(std::size_t pad_index) const;
+    Common::UUID GetHostUUID(const std::string& host) const;
 
     Common::Input::ButtonNames GetUIButtonName(const Common::ParamPackage& params) const;
 
diff --git a/src/input_common/input_engine.h b/src/input_common/input_engine.h
index 390581c94..fe2faee5a 100644
--- a/src/input_common/input_engine.h
+++ b/src/input_common/input_engine.h
@@ -16,7 +16,7 @@
 
 // Pad Identifier of data source
 struct PadIdentifier {
-    Common::UUID guid{};
+    Common::UUID guid{Common::INVALID_UUID};
     std::size_t port{};
     std::size_t pad{};
 
@@ -89,7 +89,7 @@ struct UpdateCallback {
 
 // Triggered if data changed on the controller and the engine is on configuring mode
 struct MappingCallback {
-    std::function<void(MappingData)> on_data;
+    std::function<void(const MappingData&)> on_data;
 };
 
 // Input Identifier of data source
diff --git a/src/input_common/input_mapping.cpp b/src/input_common/input_mapping.cpp
index 475257f42..a7a6ad8c2 100644
--- a/src/input_common/input_mapping.cpp
+++ b/src/input_common/input_mapping.cpp
@@ -2,14 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included
 
-#include "common/common_types.h"
 #include "common/settings.h"
 #include "input_common/input_engine.h"
 #include "input_common/input_mapping.h"
 
 namespace InputCommon {
 
-MappingFactory::MappingFactory() {}
+MappingFactory::MappingFactory() = default;
 
 void MappingFactory::BeginMapping(Polling::InputType type) {
     is_enabled = true;
@@ -19,7 +18,7 @@ void MappingFactory::BeginMapping(Polling::InputType type) {
     second_axis = -1;
 }
 
-[[nodiscard]] const Common::ParamPackage MappingFactory::GetNextInput() {
+Common::ParamPackage MappingFactory::GetNextInput() {
     Common::ParamPackage input;
     input_queue.Pop(input);
     return input;
@@ -57,7 +56,7 @@ void MappingFactory::StopMapping() {
 void MappingFactory::RegisterButton(const MappingData& data) {
     Common::ParamPackage new_input;
     new_input.Set("engine", data.engine);
-    if (data.pad.guid != Common::UUID{}) {
+    if (data.pad.guid.IsValid()) {
         new_input.Set("guid", data.pad.guid.Format());
     }
     new_input.Set("port", static_cast<int>(data.pad.port));
@@ -93,7 +92,7 @@ void MappingFactory::RegisterButton(const MappingData& data) {
 void MappingFactory::RegisterStick(const MappingData& data) {
     Common::ParamPackage new_input;
     new_input.Set("engine", data.engine);
-    if (data.pad.guid != Common::UUID{}) {
+    if (data.pad.guid.IsValid()) {
         new_input.Set("guid", data.pad.guid.Format());
     }
     new_input.Set("port", static_cast<int>(data.pad.port));
@@ -138,7 +137,7 @@ void MappingFactory::RegisterStick(const MappingData& data) {
 void MappingFactory::RegisterMotion(const MappingData& data) {
     Common::ParamPackage new_input;
     new_input.Set("engine", data.engine);
-    if (data.pad.guid != Common::UUID{}) {
+    if (data.pad.guid.IsValid()) {
         new_input.Set("guid", data.pad.guid.Format());
     }
     new_input.Set("port", static_cast<int>(data.pad.port));
diff --git a/src/input_common/input_mapping.h b/src/input_common/input_mapping.h
index 93564b5f8..e0dfbc7ad 100644
--- a/src/input_common/input_mapping.h
+++ b/src/input_common/input_mapping.h
@@ -3,8 +3,14 @@
 // Refer to the license.txt file included
 
 #pragma once
+
+#include "common/param_package.h"
 #include "common/threadsafe_queue.h"
 
+namespace InputCommon::Polling {
+enum class InputType;
+}
+
 namespace InputCommon {
 class InputEngine;
 struct MappingData;
@@ -20,7 +26,7 @@ public:
     void BeginMapping(Polling::InputType type);
 
     /// Returns an input event with mapping information from the input_queue
-    [[nodiscard]] const Common::ParamPackage GetNextInput();
+    [[nodiscard]] Common::ParamPackage GetNextInput();
 
     /**
      * Registers mapping input data from the driver
diff --git a/src/input_common/input_poller.cpp b/src/input_common/input_poller.cpp
index 7b370335f..2f3c0735a 100644
--- a/src/input_common/input_poller.cpp
+++ b/src/input_common/input_poller.cpp
@@ -504,9 +504,10 @@ private:
 
 class InputFromMotion final : public Common::Input::InputDevice {
 public:
-    explicit InputFromMotion(PadIdentifier identifier_, int motion_sensor_,
+    explicit InputFromMotion(PadIdentifier identifier_, int motion_sensor_, float gyro_threshold_,
                              InputEngine* input_engine_)
-        : identifier(identifier_), motion_sensor(motion_sensor_), input_engine(input_engine_) {
+        : identifier(identifier_), motion_sensor(motion_sensor_), gyro_threshold(gyro_threshold_),
+          input_engine(input_engine_) {
         UpdateCallback engine_callback{[this]() { OnChange(); }};
         const InputIdentifier input_identifier{
             .identifier = identifier,
@@ -525,8 +526,9 @@ public:
         const auto basic_motion = input_engine->GetMotion(identifier, motion_sensor);
         Common::Input::MotionStatus status{};
         const Common::Input::AnalogProperties properties = {
-            .deadzone = 0.001f,
+            .deadzone = 0.0f,
             .range = 1.0f,
+            .threshold = gyro_threshold,
             .offset = 0.0f,
         };
         status.accel.x = {.raw_value = basic_motion.accel_x, .properties = properties};
@@ -551,6 +553,7 @@ public:
 private:
     const PadIdentifier identifier;
     const int motion_sensor;
+    const float gyro_threshold;
     int callback_key;
     InputEngine* input_engine;
 };
@@ -873,9 +876,11 @@ std::unique_ptr<Common::Input::InputDevice> InputFactory::CreateMotionDevice(
 
     if (params.Has("motion")) {
         const auto motion_sensor = params.Get("motion", 0);
+        const auto gyro_threshold = params.Get("threshold", 0.007f);
         input_engine->PreSetController(identifier);
         input_engine->PreSetMotion(identifier, motion_sensor);
-        return std::make_unique<InputFromMotion>(identifier, motion_sensor, input_engine.get());
+        return std::make_unique<InputFromMotion>(identifier, motion_sensor, gyro_threshold,
+                                                 input_engine.get());
     }
 
     const auto deadzone = std::clamp(params.Get("deadzone", 0.15f), 0.0f, 1.0f);
diff --git a/src/input_common/main.cpp b/src/input_common/main.cpp
index 940744c5f..a4d7ed645 100644
--- a/src/input_common/main.cpp
+++ b/src/input_common/main.cpp
@@ -27,7 +27,7 @@ namespace InputCommon {
 struct InputSubsystem::Impl {
     void Initialize() {
         mapping_factory = std::make_shared<MappingFactory>();
-        MappingCallback mapping_callback{[this](MappingData data) { RegisterInput(data); }};
+        MappingCallback mapping_callback{[this](const MappingData& data) { RegisterInput(data); }};
 
         keyboard = std::make_shared<Keyboard>("keyboard");
         keyboard->SetMappingCallback(mapping_callback);
@@ -284,7 +284,7 @@ struct InputSubsystem::Impl {
 #endif
     }
 
-    void RegisterInput(MappingData data) {
+    void RegisterInput(const MappingData& data) {
         mapping_factory->RegisterInput(data);
     }
 
@@ -394,7 +394,7 @@ void InputSubsystem::BeginMapping(Polling::InputType type) {
     impl->mapping_factory->BeginMapping(type);
 }
 
-const Common::ParamPackage InputSubsystem::GetNextInput() const {
+Common::ParamPackage InputSubsystem::GetNextInput() const {
     return impl->mapping_factory->GetNextInput();
 }
 
diff --git a/src/input_common/main.h b/src/input_common/main.h
index c6f97f691..baf107e0f 100644
--- a/src/input_common/main.h
+++ b/src/input_common/main.h
@@ -126,7 +126,7 @@ public:
     void BeginMapping(Polling::InputType type);
 
     /// Returns an input event with mapping information.
-    [[nodiscard]] const Common::ParamPackage GetNextInput() const;
+    [[nodiscard]] Common::ParamPackage GetNextInput() const;
 
     /// Stop polling from all backends.
     void StopMapping() const;
diff --git a/src/shader_recompiler/backend/spirv/emit_spirv.cpp b/src/shader_recompiler/backend/spirv/emit_spirv.cpp
index 50918317f..08b3a81ce 100644
--- a/src/shader_recompiler/backend/spirv/emit_spirv.cpp
+++ b/src/shader_recompiler/backend/spirv/emit_spirv.cpp
@@ -387,6 +387,14 @@ void SetupSignedNanCapabilities(const Profile& profile, const IR::Program& progr
     }
 }
 
+void SetupTransformFeedbackCapabilities(EmitContext& ctx, Id main_func) {
+    if (ctx.runtime_info.xfb_varyings.empty()) {
+        return;
+    }
+    ctx.AddCapability(spv::Capability::TransformFeedback);
+    ctx.AddExecutionMode(main_func, spv::ExecutionMode::Xfb);
+}
+
 void SetupCapabilities(const Profile& profile, const Info& info, EmitContext& ctx) {
     if (info.uses_sampled_1d) {
         ctx.AddCapability(spv::Capability::Sampled1D);
@@ -442,9 +450,6 @@ void SetupCapabilities(const Profile& profile, const Info& info, EmitContext& ct
     if (info.uses_sample_id) {
         ctx.AddCapability(spv::Capability::SampleRateShading);
     }
-    if (!ctx.runtime_info.xfb_varyings.empty()) {
-        ctx.AddCapability(spv::Capability::TransformFeedback);
-    }
     if (info.uses_derivatives) {
         ctx.AddCapability(spv::Capability::DerivativeControl);
     }
@@ -484,6 +489,7 @@ std::vector<u32> EmitSPIRV(const Profile& profile, const RuntimeInfo& runtime_in
         SetupSignedNanCapabilities(profile, program, ctx, main);
     }
     SetupCapabilities(profile, program.info, ctx);
+    SetupTransformFeedbackCapabilities(ctx, main);
     PatchPhiNodes(program, ctx);
     return ctx.Assemble();
 }
diff --git a/src/shader_recompiler/backend/spirv/emit_spirv_atomic.cpp b/src/shader_recompiler/backend/spirv/emit_spirv_atomic.cpp
index 0d37b405c..46ba52a25 100644
--- a/src/shader_recompiler/backend/spirv/emit_spirv_atomic.cpp
+++ b/src/shader_recompiler/backend/spirv/emit_spirv_atomic.cpp
@@ -74,7 +74,7 @@ Id StorageAtomicU64(EmitContext& ctx, const IR::Value& binding, const IR::Value&
         const auto [scope, semantics]{AtomicArgs(ctx)};
         return (ctx.*atomic_func)(ctx.U64, pointer, scope, semantics, value);
     }
-    LOG_ERROR(Shader_SPIRV, "Int64 atomics not supported, fallback to non-atomic");
+    LOG_WARNING(Shader_SPIRV, "Int64 atomics not supported, fallback to non-atomic");
     const Id pointer{StoragePointer(ctx, ctx.storage_types.U32x2, &StorageDefinitions::U32x2,
                                     binding, offset, sizeof(u32[2]))};
     const Id original_value{ctx.OpBitcast(ctx.U64, ctx.OpLoad(ctx.U32[2], pointer))};
@@ -267,7 +267,7 @@ Id EmitStorageAtomicExchange64(EmitContext& ctx, const IR::Value& binding, const
         const auto [scope, semantics]{AtomicArgs(ctx)};
         return ctx.OpAtomicExchange(ctx.U64, pointer, scope, semantics, value);
     }
-    LOG_ERROR(Shader_SPIRV, "Int64 atomics not supported, fallback to non-atomic");
+    LOG_WARNING(Shader_SPIRV, "Int64 atomics not supported, fallback to non-atomic");
     const Id pointer{StoragePointer(ctx, ctx.storage_types.U32x2, &StorageDefinitions::U32x2,
                                     binding, offset, sizeof(u32[2]))};
     const Id original{ctx.OpBitcast(ctx.U64, ctx.OpLoad(ctx.U32[2], pointer))};
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/video_minimum_maximum.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/video_minimum_maximum.cpp
index 78869601f..4851b0b8d 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/video_minimum_maximum.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/video_minimum_maximum.cpp
@@ -57,16 +57,6 @@ void TranslatorVisitor::VMNMX(u64 insn) {
     if (vmnmx.sat != 0) {
         throw NotImplementedException("VMNMX SAT");
     }
-    // Selectors were shown to default to 2 in unit tests
-    if (vmnmx.src_a_selector != 2) {
-        throw NotImplementedException("VMNMX Selector {}", vmnmx.src_a_selector.Value());
-    }
-    if (vmnmx.src_b_selector != 2) {
-        throw NotImplementedException("VMNMX Selector {}", vmnmx.src_b_selector.Value());
-    }
-    if (vmnmx.src_a_width != VideoWidth::Word) {
-        throw NotImplementedException("VMNMX Source Width {}", vmnmx.src_a_width.Value());
-    }
 
     const bool is_b_imm{vmnmx.is_src_b_reg == 0};
     const IR::U32 src_a{GetReg8(insn)};
@@ -76,10 +66,14 @@ void TranslatorVisitor::VMNMX(u64 insn) {
     const VideoWidth a_width{vmnmx.src_a_width};
     const VideoWidth b_width{GetVideoSourceWidth(vmnmx.src_b_width, is_b_imm)};
 
+    const u32 a_selector{static_cast<u32>(vmnmx.src_a_selector)};
+    // Immediate values can't have a selector
+    const u32 b_selector{is_b_imm ? 0U : static_cast<u32>(vmnmx.src_b_selector)};
+
     const bool src_a_signed{vmnmx.src_a_sign != 0};
     const bool src_b_signed{vmnmx.src_b_sign != 0};
-    const IR::U32 op_a{ExtractVideoOperandValue(ir, src_a, a_width, 0, src_a_signed)};
-    const IR::U32 op_b{ExtractVideoOperandValue(ir, src_b, b_width, 0, src_b_signed)};
+    const IR::U32 op_a{ExtractVideoOperandValue(ir, src_a, a_width, a_selector, src_a_signed)};
+    const IR::U32 op_b{ExtractVideoOperandValue(ir, src_b, b_width, b_selector, src_b_signed)};
 
     // First operation's sign is only dependent on operand b's sign
     const bool op_1_signed{src_b_signed};
diff --git a/src/shader_recompiler/frontend/maxwell/translate_program.h b/src/shader_recompiler/frontend/maxwell/translate_program.h
index cd535f20d..eac83da9d 100644
--- a/src/shader_recompiler/frontend/maxwell/translate_program.h
+++ b/src/shader_recompiler/frontend/maxwell/translate_program.h
@@ -21,7 +21,6 @@ namespace Shader::Maxwell {
 [[nodiscard]] IR::Program MergeDualVertexPrograms(IR::Program& vertex_a, IR::Program& vertex_b,
                                                   Environment& env_vertex_b);
 
-[[nodiscard]] void ConvertLegacyToGeneric(IR::Program& program,
-                                          const Shader::RuntimeInfo& runtime_info);
+void ConvertLegacyToGeneric(IR::Program& program, const RuntimeInfo& runtime_info);
 
 } // namespace Shader::Maxwell
diff --git a/src/shader_recompiler/ir_opt/collect_shader_info_pass.cpp b/src/shader_recompiler/ir_opt/collect_shader_info_pass.cpp
index a78c469be..b6a20f904 100644
--- a/src/shader_recompiler/ir_opt/collect_shader_info_pass.cpp
+++ b/src/shader_recompiler/ir_opt/collect_shader_info_pass.cpp
@@ -688,7 +688,7 @@ void VisitUsages(Info& info, IR::Inst& inst) {
     case IR::Opcode::StorageAtomicAnd64:
     case IR::Opcode::StorageAtomicOr64:
     case IR::Opcode::StorageAtomicXor64:
-        info.used_storage_buffer_types |= IR::Type::U64;
+        info.used_storage_buffer_types |= IR::Type::U64 | IR::Type::U32x2;
         info.uses_int64_bit_atomics = true;
         break;
     case IR::Opcode::BindlessImageAtomicIAdd32:
diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h
index 43bed63ac..fa26eb8b0 100644
--- a/src/video_core/buffer_cache/buffer_cache.h
+++ b/src/video_core/buffer_cache/buffer_cache.h
@@ -131,6 +131,8 @@ public:
 
     void DownloadMemory(VAddr cpu_addr, u64 size);
 
+    bool InlineMemory(VAddr dest_address, size_t copy_size, std::span<u8> inlined_buffer);
+
     void BindGraphicsUniformBuffer(size_t stage, u32 index, GPUVAddr gpu_addr, u32 size);
 
     void DisableGraphicsUniformBuffer(size_t stage, u32 index);
@@ -808,6 +810,8 @@ void BufferCache<P>::CommitAsyncFlushesHigh() {
         return;
     }
     MICROPROFILE_SCOPE(GPU_DownloadMemory);
+    const bool is_accuracy_normal =
+        Settings::values.gpu_accuracy.GetValue() == Settings::GPUAccuracy::Normal;
 
     boost::container::small_vector<std::pair<BufferCopy, BufferId>, 1> downloads;
     u64 total_size_bytes = 0;
@@ -819,6 +823,9 @@ void BufferCache<P>::CommitAsyncFlushesHigh() {
             ForEachBufferInRange(cpu_addr, size, [&](BufferId buffer_id, Buffer& buffer) {
                 buffer.ForEachDownloadRangeAndClear(
                     cpu_addr, size, [&](u64 range_offset, u64 range_size) {
+                        if (is_accuracy_normal) {
+                            return;
+                        }
                         const VAddr buffer_addr = buffer.CpuAddr();
                         const auto add_download = [&](VAddr start, VAddr end) {
                             const u64 new_offset = start - buffer_addr;
@@ -1417,10 +1424,8 @@ void BufferCache<P>::MarkWrittenBuffer(BufferId buffer_id, VAddr cpu_addr, u32 s
     const IntervalType base_interval{cpu_addr, cpu_addr + size};
     common_ranges.add(base_interval);
 
-    const bool is_accuracy_high =
-        Settings::values.gpu_accuracy.GetValue() == Settings::GPUAccuracy::High;
     const bool is_async = Settings::values.use_asynchronous_gpu_emulation.GetValue();
-    if (!is_async && !is_accuracy_high) {
+    if (!is_async) {
         return;
     }
     uncommitted_ranges.add(base_interval);
@@ -1474,6 +1479,8 @@ typename BufferCache<P>::OverlapResult BufferCache<P>::ResolveOverlaps(VAddr cpu
             // When this memory region has been joined a bunch of times, we assume it's being used
             // as a stream buffer. Increase the size to skip constantly recreating buffers.
             has_stream_leap = true;
+            begin -= PAGE_SIZE * 256;
+            cpu_addr = begin;
             end += PAGE_SIZE * 256;
         }
     }
@@ -1642,6 +1649,42 @@ void BufferCache<P>::MappedUploadMemory(Buffer& buffer, u64 total_size_bytes,
 }
 
 template <class P>
+bool BufferCache<P>::InlineMemory(VAddr dest_address, size_t copy_size,
+                                  std::span<u8> inlined_buffer) {
+    const bool is_dirty = IsRegionRegistered(dest_address, copy_size);
+    if (!is_dirty) {
+        return false;
+    }
+    if (!IsRegionGpuModified(dest_address, copy_size)) {
+        return false;
+    }
+
+    const IntervalType subtract_interval{dest_address, dest_address + copy_size};
+    ClearDownload(subtract_interval);
+    common_ranges.subtract(subtract_interval);
+
+    BufferId buffer_id = FindBuffer(dest_address, static_cast<u32>(copy_size));
+    auto& buffer = slot_buffers[buffer_id];
+    SynchronizeBuffer(buffer, dest_address, static_cast<u32>(copy_size));
+
+    if constexpr (USE_MEMORY_MAPS) {
+        std::array copies{BufferCopy{
+            .src_offset = 0,
+            .dst_offset = buffer.Offset(dest_address),
+            .size = copy_size,
+        }};
+        auto upload_staging = runtime.UploadStagingBuffer(copy_size);
+        u8* const src_pointer = upload_staging.mapped_span.data();
+        std::memcpy(src_pointer, inlined_buffer.data(), copy_size);
+        runtime.CopyBuffer(buffer, upload_staging.buffer, copies);
+    } else {
+        buffer.ImmediateUpload(buffer.Offset(dest_address), inlined_buffer.first(copy_size));
+    }
+
+    return true;
+}
+
+template <class P>
 void BufferCache<P>::DownloadBufferMemory(Buffer& buffer) {
     DownloadBufferMemory(buffer, buffer.CpuAddr(), buffer.SizeBytes());
 }
diff --git a/src/video_core/engines/engine_upload.cpp b/src/video_core/engines/engine_upload.cpp
index 71d7e1473..351b110fe 100644
--- a/src/video_core/engines/engine_upload.cpp
+++ b/src/video_core/engines/engine_upload.cpp
@@ -7,6 +7,7 @@
 #include "common/assert.h"
 #include "video_core/engines/engine_upload.h"
 #include "video_core/memory_manager.h"
+#include "video_core/rasterizer_interface.h"
 #include "video_core/textures/decoders.h"
 
 namespace Tegra::Engines::Upload {
@@ -16,6 +17,10 @@ State::State(MemoryManager& memory_manager_, Registers& regs_)
 
 State::~State() = default;
 
+void State::BindRasterizer(VideoCore::RasterizerInterface* rasterizer_) {
+    rasterizer = rasterizer_;
+}
+
 void State::ProcessExec(const bool is_linear_) {
     write_offset = 0;
     copy_size = regs.line_length_in * regs.line_count;
@@ -32,7 +37,7 @@ void State::ProcessData(const u32 data, const bool is_last_call) {
     }
     const GPUVAddr address{regs.dest.Address()};
     if (is_linear) {
-        memory_manager.WriteBlock(address, inner_buffer.data(), copy_size);
+        rasterizer->AccelerateInlineToMemory(address, copy_size, inner_buffer);
     } else {
         UNIMPLEMENTED_IF(regs.dest.z != 0);
         UNIMPLEMENTED_IF(regs.dest.depth != 1);
diff --git a/src/video_core/engines/engine_upload.h b/src/video_core/engines/engine_upload.h
index 1c7f1effa..c9c5ec8c3 100644
--- a/src/video_core/engines/engine_upload.h
+++ b/src/video_core/engines/engine_upload.h
@@ -12,6 +12,10 @@ namespace Tegra {
 class MemoryManager;
 }
 
+namespace VideoCore {
+class RasterizerInterface;
+}
+
 namespace Tegra::Engines::Upload {
 
 struct Registers {
@@ -60,6 +64,9 @@ public:
     void ProcessExec(bool is_linear_);
     void ProcessData(u32 data, bool is_last_call);
 
+    /// Binds a rasterizer to this engine.
+    void BindRasterizer(VideoCore::RasterizerInterface* rasterizer);
+
 private:
     u32 write_offset = 0;
     u32 copy_size = 0;
@@ -68,6 +75,7 @@ private:
     bool is_linear = false;
     Registers& regs;
     MemoryManager& memory_manager;
+    VideoCore::RasterizerInterface* rasterizer = nullptr;
 };
 
 } // namespace Tegra::Engines::Upload
diff --git a/src/video_core/engines/kepler_compute.cpp b/src/video_core/engines/kepler_compute.cpp
index 492b4c5a3..5a1c12076 100644
--- a/src/video_core/engines/kepler_compute.cpp
+++ b/src/video_core/engines/kepler_compute.cpp
@@ -22,6 +22,7 @@ KeplerCompute::~KeplerCompute() = default;
 
 void KeplerCompute::BindRasterizer(VideoCore::RasterizerInterface* rasterizer_) {
     rasterizer = rasterizer_;
+    upload_state.BindRasterizer(rasterizer);
 }
 
 void KeplerCompute::CallMethod(u32 method, u32 method_argument, bool is_last_call) {
diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp
index 560551157..8aed16caa 100644
--- a/src/video_core/engines/kepler_memory.cpp
+++ b/src/video_core/engines/kepler_memory.cpp
@@ -19,6 +19,10 @@ KeplerMemory::KeplerMemory(Core::System& system_, MemoryManager& memory_manager)
 
 KeplerMemory::~KeplerMemory() = default;
 
+void KeplerMemory::BindRasterizer(VideoCore::RasterizerInterface* rasterizer_) {
+    upload_state.BindRasterizer(rasterizer_);
+}
+
 void KeplerMemory::CallMethod(u32 method, u32 method_argument, bool is_last_call) {
     ASSERT_MSG(method < Regs::NUM_REGS,
                "Invalid KeplerMemory register, increase the size of the Regs structure");
diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h
index 0d8ea09a9..949e2fae1 100644
--- a/src/video_core/engines/kepler_memory.h
+++ b/src/video_core/engines/kepler_memory.h
@@ -22,6 +22,10 @@ namespace Tegra {
 class MemoryManager;
 }
 
+namespace VideoCore {
+class RasterizerInterface;
+}
+
 namespace Tegra::Engines {
 
 /**
@@ -38,6 +42,9 @@ public:
     explicit KeplerMemory(Core::System& system_, MemoryManager& memory_manager);
     ~KeplerMemory() override;
 
+    /// Binds a rasterizer to this engine.
+    void BindRasterizer(VideoCore::RasterizerInterface* rasterizer);
+
     /// Write the value to the register identified by method.
     void CallMethod(u32 method, u32 method_argument, bool is_last_call) override;
 
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index b18b8a02a..5d6d217bb 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -31,6 +31,7 @@ Maxwell3D::~Maxwell3D() = default;
 
 void Maxwell3D::BindRasterizer(VideoCore::RasterizerInterface* rasterizer_) {
     rasterizer = rasterizer_;
+    upload_state.BindRasterizer(rasterizer_);
 }
 
 void Maxwell3D::InitializeRegisterDefaults() {
@@ -240,7 +241,7 @@ void Maxwell3D::CallMacroMethod(u32 method, const std::vector<u32>& parameters)
         ((method - MacroRegistersStart) >> 1) % static_cast<u32>(macro_positions.size());
 
     // Execute the current macro.
-    macro_engine->Execute(*this, macro_positions[entry], parameters);
+    macro_engine->Execute(macro_positions[entry], parameters);
     if (mme_draw.current_mode != MMEDrawMode::Undefined) {
         FlushMMEInlineDraw();
     }
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index f22342dfb..dc9df6c8b 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -1557,7 +1557,8 @@ private:
 
     static constexpr u32 null_cb_data = 0xFFFFFFFF;
     struct CBDataState {
-        std::array<std::array<u32, 0x4000>, 16> buffer;
+        static constexpr size_t inline_size = 0x4000;
+        std::array<std::array<u32, inline_size>, 16> buffer;
         u32 current{null_cb_data};
         u32 id{null_cb_data};
         u32 start_pos{};
diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp
index 705765c99..ba9ba082f 100644
--- a/src/video_core/gpu.cpp
+++ b/src/video_core/gpu.cpp
@@ -59,6 +59,7 @@ struct GPU::Impl {
         maxwell_3d->BindRasterizer(rasterizer);
         fermi_2d->BindRasterizer(rasterizer);
         kepler_compute->BindRasterizer(rasterizer);
+        kepler_memory->BindRasterizer(rasterizer);
         maxwell_dma->BindRasterizer(rasterizer);
     }
 
@@ -502,8 +503,13 @@ struct GPU::Impl {
         case BufferMethods::SemaphoreAddressHigh:
         case BufferMethods::SemaphoreAddressLow:
         case BufferMethods::SemaphoreSequence:
+            break;
         case BufferMethods::UnkCacheFlush:
+            rasterizer->SyncGuestHost();
+            break;
         case BufferMethods::WrcacheFlush:
+            rasterizer->SignalReference();
+            break;
         case BufferMethods::FenceValue:
             break;
         case BufferMethods::RefCnt:
@@ -513,7 +519,7 @@ struct GPU::Impl {
             ProcessFenceActionMethod();
             break;
         case BufferMethods::WaitForInterrupt:
-            ProcessWaitForInterruptMethod();
+            rasterizer->WaitForIdle();
             break;
         case BufferMethods::SemaphoreTrigger: {
             ProcessSemaphoreTriggerMethod();
diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h
index 3188b83ed..26b8ea233 100644
--- a/src/video_core/gpu.h
+++ b/src/video_core/gpu.h
@@ -12,9 +12,6 @@
 #include "video_core/framebuffer_config.h"
 
 namespace Core {
-namespace Frontend {
-class EmuWindow;
-}
 class System;
 } // namespace Core
 
@@ -25,7 +22,6 @@ class ShaderNotify;
 
 namespace Tegra {
 class DmaPusher;
-class CDmaPusher;
 struct CommandList;
 
 enum class RenderTargetFormat : u32 {
@@ -88,15 +84,9 @@ enum class DepthFormat : u32 {
     D32_FLOAT_S8X24_UINT = 0x19,
 };
 
-struct CommandListHeader;
-class DebugContext;
-
 namespace Engines {
-class Fermi2D;
 class Maxwell3D;
-class MaxwellDMA;
 class KeplerCompute;
-class KeplerMemory;
 } // namespace Engines
 
 enum class EngineID {
@@ -190,12 +180,6 @@ public:
     /// Returns a const reference to the GPU DMA pusher.
     [[nodiscard]] const Tegra::DmaPusher& DmaPusher() const;
 
-    /// Returns a reference to the GPU CDMA pusher.
-    [[nodiscard]] Tegra::CDmaPusher& CDmaPusher();
-
-    /// Returns a const reference to the GPU CDMA pusher.
-    [[nodiscard]] const Tegra::CDmaPusher& CDmaPusher() const;
-
     /// Returns a reference to the underlying renderer.
     [[nodiscard]] VideoCore::RendererBase& Renderer();
 
diff --git a/src/video_core/macro/macro.cpp b/src/video_core/macro/macro.cpp
index d7fabe605..0aeda4ce8 100644
--- a/src/video_core/macro/macro.cpp
+++ b/src/video_core/macro/macro.cpp
@@ -2,12 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cstring>
 #include <optional>
+
 #include <boost/container_hash/hash.hpp>
+
 #include "common/assert.h"
-#include "common/logging/log.h"
 #include "common/settings.h"
-#include "video_core/engines/maxwell_3d.h"
 #include "video_core/macro/macro.h"
 #include "video_core/macro/macro_hle.h"
 #include "video_core/macro/macro_interpreter.h"
@@ -24,8 +25,7 @@ void MacroEngine::AddCode(u32 method, u32 data) {
     uploaded_macro_code[method].push_back(data);
 }
 
-void MacroEngine::Execute(Engines::Maxwell3D& maxwell3d, u32 method,
-                          const std::vector<u32>& parameters) {
+void MacroEngine::Execute(u32 method, const std::vector<u32>& parameters) {
     auto compiled_macro = macro_cache.find(method);
     if (compiled_macro != macro_cache.end()) {
         const auto& cache_info = compiled_macro->second;
@@ -66,10 +66,9 @@ void MacroEngine::Execute(Engines::Maxwell3D& maxwell3d, u32 method,
             cache_info.lle_program = Compile(code);
         }
 
-        auto hle_program = hle_macros->GetHLEProgram(cache_info.hash);
-        if (hle_program.has_value()) {
+        if (auto hle_program = hle_macros->GetHLEProgram(cache_info.hash)) {
             cache_info.has_hle_program = true;
-            cache_info.hle_program = std::move(hle_program.value());
+            cache_info.hle_program = std::move(hle_program);
             cache_info.hle_program->Execute(parameters, method);
         } else {
             cache_info.lle_program->Execute(parameters, method);
diff --git a/src/video_core/macro/macro.h b/src/video_core/macro/macro.h
index 31ee3440a..7aaa49286 100644
--- a/src/video_core/macro/macro.h
+++ b/src/video_core/macro/macro.h
@@ -119,7 +119,7 @@ public:
     void AddCode(u32 method, u32 data);
 
     // Compiles the macro if its not in the cache, and executes the compiled macro
-    void Execute(Engines::Maxwell3D& maxwell3d, u32 method, const std::vector<u32>& parameters);
+    void Execute(u32 method, const std::vector<u32>& parameters);
 
 protected:
     virtual std::unique_ptr<CachedMacro> Compile(const std::vector<u32>& code) = 0;
diff --git a/src/video_core/macro/macro_hle.cpp b/src/video_core/macro/macro_hle.cpp
index 70ac7c620..900ad23c9 100644
--- a/src/video_core/macro/macro_hle.cpp
+++ b/src/video_core/macro/macro_hle.cpp
@@ -5,12 +5,15 @@
 #include <array>
 #include <vector>
 #include "video_core/engines/maxwell_3d.h"
+#include "video_core/macro/macro.h"
 #include "video_core/macro/macro_hle.h"
 #include "video_core/rasterizer_interface.h"
 
 namespace Tegra {
-
 namespace {
+
+using HLEFunction = void (*)(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters);
+
 // HLE'd functions
 void HLE_771BB18C62444DA0(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) {
     const u32 instance_count = parameters[2] & maxwell3d.GetRegisterValue(0xD1B);
@@ -77,7 +80,6 @@ void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d, const std::vector<u32>&
     maxwell3d.CallMethodFromMME(0x8e5, 0x0);
     maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined;
 }
-} // Anonymous namespace
 
 constexpr std::array<std::pair<u64, HLEFunction>, 3> hle_funcs{{
     {0x771BB18C62444DA0, &HLE_771BB18C62444DA0},
@@ -85,25 +87,31 @@ constexpr std::array<std::pair<u64, HLEFunction>, 3> hle_funcs{{
     {0x0217920100488FF7, &HLE_0217920100488FF7},
 }};
 
+class HLEMacroImpl final : public CachedMacro {
+public:
+    explicit HLEMacroImpl(Engines::Maxwell3D& maxwell3d_, HLEFunction func_)
+        : maxwell3d{maxwell3d_}, func{func_} {}
+
+    void Execute(const std::vector<u32>& parameters, u32 method) override {
+        func(maxwell3d, parameters);
+    }
+
+private:
+    Engines::Maxwell3D& maxwell3d;
+    HLEFunction func;
+};
+} // Anonymous namespace
+
 HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d_) : maxwell3d{maxwell3d_} {}
 HLEMacro::~HLEMacro() = default;
 
-std::optional<std::unique_ptr<CachedMacro>> HLEMacro::GetHLEProgram(u64 hash) const {
+std::unique_ptr<CachedMacro> HLEMacro::GetHLEProgram(u64 hash) const {
     const auto it = std::find_if(hle_funcs.cbegin(), hle_funcs.cend(),
                                  [hash](const auto& pair) { return pair.first == hash; });
     if (it == hle_funcs.end()) {
-        return std::nullopt;
+        return nullptr;
     }
     return std::make_unique<HLEMacroImpl>(maxwell3d, it->second);
 }
 
-HLEMacroImpl::~HLEMacroImpl() = default;
-
-HLEMacroImpl::HLEMacroImpl(Engines::Maxwell3D& maxwell3d_, HLEFunction func_)
-    : maxwell3d{maxwell3d_}, func{func_} {}
-
-void HLEMacroImpl::Execute(const std::vector<u32>& parameters, u32 method) {
-    func(maxwell3d, parameters);
-}
-
 } // namespace Tegra
diff --git a/src/video_core/macro/macro_hle.h b/src/video_core/macro/macro_hle.h
index cb3bd1600..b86ba84a1 100644
--- a/src/video_core/macro/macro_hle.h
+++ b/src/video_core/macro/macro_hle.h
@@ -5,10 +5,7 @@
 #pragma once
 
 #include <memory>
-#include <optional>
-#include <vector>
 #include "common/common_types.h"
-#include "video_core/macro/macro.h"
 
 namespace Tegra {
 
@@ -16,29 +13,17 @@ namespace Engines {
 class Maxwell3D;
 }
 
-using HLEFunction = void (*)(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters);
-
 class HLEMacro {
 public:
     explicit HLEMacro(Engines::Maxwell3D& maxwell3d_);
     ~HLEMacro();
 
-    std::optional<std::unique_ptr<CachedMacro>> GetHLEProgram(u64 hash) const;
-
-private:
-    Engines::Maxwell3D& maxwell3d;
-};
-
-class HLEMacroImpl : public CachedMacro {
-public:
-    explicit HLEMacroImpl(Engines::Maxwell3D& maxwell3d, HLEFunction func);
-    ~HLEMacroImpl();
-
-    void Execute(const std::vector<u32>& parameters, u32 method) override;
+    // Allocates and returns a cached macro if the hash matches a known function.
+    // Returns nullptr otherwise.
+    [[nodiscard]] std::unique_ptr<CachedMacro> GetHLEProgram(u64 hash) const;
 
 private:
     Engines::Maxwell3D& maxwell3d;
-    HLEFunction func;
 };
 
 } // namespace Tegra
diff --git a/src/video_core/macro/macro_interpreter.cpp b/src/video_core/macro/macro_interpreter.cpp
index 8da26fd59..fba755448 100644
--- a/src/video_core/macro/macro_interpreter.cpp
+++ b/src/video_core/macro/macro_interpreter.cpp
@@ -2,6 +2,9 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <array>
+#include <optional>
+
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "common/microprofile.h"
@@ -11,16 +14,81 @@
 MICROPROFILE_DEFINE(MacroInterp, "GPU", "Execute macro interpreter", MP_RGB(128, 128, 192));
 
 namespace Tegra {
-MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d_)
-    : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {}
+namespace {
+class MacroInterpreterImpl final : public CachedMacro {
+public:
+    explicit MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_)
+        : maxwell3d{maxwell3d_}, code{code_} {}
 
-std::unique_ptr<CachedMacro> MacroInterpreter::Compile(const std::vector<u32>& code) {
-    return std::make_unique<MacroInterpreterImpl>(maxwell3d, code);
-}
+    void Execute(const std::vector<u32>& params, u32 method) override;
+
+private:
+    /// Resets the execution engine state, zeroing registers, etc.
+    void Reset();
+
+    /**
+     * Executes a single macro instruction located at the current program counter. Returns whether
+     * the interpreter should keep running.
+     *
+     * @param is_delay_slot Whether the current step is being executed due to a delay slot in a
+     *                      previous instruction.
+     */
+    bool Step(bool is_delay_slot);
+
+    /// Calculates the result of an ALU operation. src_a OP src_b;
+    u32 GetALUResult(Macro::ALUOperation operation, u32 src_a, u32 src_b);
+
+    /// Performs the result operation on the input result and stores it in the specified register
+    /// (if necessary).
+    void ProcessResult(Macro::ResultOperation operation, u32 reg, u32 result);
+
+    /// Evaluates the branch condition and returns whether the branch should be taken or not.
+    bool EvaluateBranchCondition(Macro::BranchCondition cond, u32 value) const;
+
+    /// Reads an opcode at the current program counter location.
+    Macro::Opcode GetOpcode() const;
+
+    /// Returns the specified register's value. Register 0 is hardcoded to always return 0.
+    u32 GetRegister(u32 register_id) const;
+
+    /// Sets the register to the input value.
+    void SetRegister(u32 register_id, u32 value);
+
+    /// Sets the method address to use for the next Send instruction.
+    void SetMethodAddress(u32 address);
 
-MacroInterpreterImpl::MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_,
-                                           const std::vector<u32>& code_)
-    : maxwell3d{maxwell3d_}, code{code_} {}
+    /// Calls a GPU Engine method with the input parameter.
+    void Send(u32 value);
+
+    /// Reads a GPU register located at the method address.
+    u32 Read(u32 method) const;
+
+    /// Returns the next parameter in the parameter queue.
+    u32 FetchParameter();
+
+    Engines::Maxwell3D& maxwell3d;
+
+    /// Current program counter
+    u32 pc{};
+    /// Program counter to execute at after the delay slot is executed.
+    std::optional<u32> delayed_pc;
+
+    /// General purpose macro registers.
+    std::array<u32, Macro::NUM_MACRO_REGISTERS> registers = {};
+
+    /// Method address to use for the next Send instruction.
+    Macro::MethodAddress method_address = {};
+
+    /// Input parameters of the current macro.
+    std::unique_ptr<u32[]> parameters;
+    std::size_t num_parameters = 0;
+    std::size_t parameters_capacity = 0;
+    /// Index of the next parameter that will be fetched by the 'parm' instruction.
+    u32 next_parameter_index = 0;
+
+    bool carry_flag = false;
+    const std::vector<u32>& code;
+};
 
 void MacroInterpreterImpl::Execute(const std::vector<u32>& params, u32 method) {
     MICROPROFILE_SCOPE(MacroInterp);
@@ -283,5 +351,13 @@ u32 MacroInterpreterImpl::FetchParameter() {
     ASSERT(next_parameter_index < num_parameters);
     return parameters[next_parameter_index++];
 }
+} // Anonymous namespace
+
+MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d_)
+    : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {}
+
+std::unique_ptr<CachedMacro> MacroInterpreter::Compile(const std::vector<u32>& code) {
+    return std::make_unique<MacroInterpreterImpl>(maxwell3d, code);
+}
 
 } // namespace Tegra
diff --git a/src/video_core/macro/macro_interpreter.h b/src/video_core/macro/macro_interpreter.h
index d50c619ce..8a9648e46 100644
--- a/src/video_core/macro/macro_interpreter.h
+++ b/src/video_core/macro/macro_interpreter.h
@@ -3,10 +3,9 @@
 // Refer to the license.txt file included.
 
 #pragma once
-#include <array>
-#include <optional>
+
 #include <vector>
-#include "common/bit_field.h"
+
 #include "common/common_types.h"
 #include "video_core/macro/macro.h"
 
@@ -26,77 +25,4 @@ private:
     Engines::Maxwell3D& maxwell3d;
 };
 
-class MacroInterpreterImpl : public CachedMacro {
-public:
-    explicit MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_);
-    void Execute(const std::vector<u32>& params, u32 method) override;
-
-private:
-    /// Resets the execution engine state, zeroing registers, etc.
-    void Reset();
-
-    /**
-     * Executes a single macro instruction located at the current program counter. Returns whether
-     * the interpreter should keep running.
-     *
-     * @param is_delay_slot Whether the current step is being executed due to a delay slot in a
-     *                      previous instruction.
-     */
-    bool Step(bool is_delay_slot);
-
-    /// Calculates the result of an ALU operation. src_a OP src_b;
-    u32 GetALUResult(Macro::ALUOperation operation, u32 src_a, u32 src_b);
-
-    /// Performs the result operation on the input result and stores it in the specified register
-    /// (if necessary).
-    void ProcessResult(Macro::ResultOperation operation, u32 reg, u32 result);
-
-    /// Evaluates the branch condition and returns whether the branch should be taken or not.
-    bool EvaluateBranchCondition(Macro::BranchCondition cond, u32 value) const;
-
-    /// Reads an opcode at the current program counter location.
-    Macro::Opcode GetOpcode() const;
-
-    /// Returns the specified register's value. Register 0 is hardcoded to always return 0.
-    u32 GetRegister(u32 register_id) const;
-
-    /// Sets the register to the input value.
-    void SetRegister(u32 register_id, u32 value);
-
-    /// Sets the method address to use for the next Send instruction.
-    void SetMethodAddress(u32 address);
-
-    /// Calls a GPU Engine method with the input parameter.
-    void Send(u32 value);
-
-    /// Reads a GPU register located at the method address.
-    u32 Read(u32 method) const;
-
-    /// Returns the next parameter in the parameter queue.
-    u32 FetchParameter();
-
-    Engines::Maxwell3D& maxwell3d;
-
-    /// Current program counter
-    u32 pc;
-    /// Program counter to execute at after the delay slot is executed.
-    std::optional<u32> delayed_pc;
-
-    /// General purpose macro registers.
-    std::array<u32, Macro::NUM_MACRO_REGISTERS> registers = {};
-
-    /// Method address to use for the next Send instruction.
-    Macro::MethodAddress method_address = {};
-
-    /// Input parameters of the current macro.
-    std::unique_ptr<u32[]> parameters;
-    std::size_t num_parameters = 0;
-    std::size_t parameters_capacity = 0;
-    /// Index of the next parameter that will be fetched by the 'parm' instruction.
-    u32 next_parameter_index = 0;
-
-    bool carry_flag = false;
-    const std::vector<u32>& code;
-};
-
 } // namespace Tegra
diff --git a/src/video_core/macro/macro_jit_x64.cpp b/src/video_core/macro/macro_jit_x64.cpp
index c6b2b2109..47b28ad16 100644
--- a/src/video_core/macro/macro_jit_x64.cpp
+++ b/src/video_core/macro/macro_jit_x64.cpp
@@ -2,9 +2,17 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <array>
+#include <bitset>
+#include <optional>
+
+#include <xbyak/xbyak.h>
+
 #include "common/assert.h"
+#include "common/bit_field.h"
 #include "common/logging/log.h"
 #include "common/microprofile.h"
+#include "common/x64/xbyak_abi.h"
 #include "common/x64/xbyak_util.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/macro/macro_interpreter.h"
@@ -14,13 +22,14 @@ MICROPROFILE_DEFINE(MacroJitCompile, "GPU", "Compile macro JIT", MP_RGB(173, 255
 MICROPROFILE_DEFINE(MacroJitExecute, "GPU", "Execute macro JIT", MP_RGB(255, 255, 0));
 
 namespace Tegra {
+namespace {
 constexpr Xbyak::Reg64 STATE = Xbyak::util::rbx;
 constexpr Xbyak::Reg32 RESULT = Xbyak::util::ebp;
 constexpr Xbyak::Reg64 PARAMETERS = Xbyak::util::r12;
 constexpr Xbyak::Reg32 METHOD_ADDRESS = Xbyak::util::r14d;
 constexpr Xbyak::Reg64 BRANCH_HOLDER = Xbyak::util::r15;
 
-static const std::bitset<32> PERSISTENT_REGISTERS = Common::X64::BuildRegSet({
+constexpr std::bitset<32> PERSISTENT_REGISTERS = Common::X64::BuildRegSet({
     STATE,
     RESULT,
     PARAMETERS,
@@ -28,19 +37,75 @@ static const std::bitset<32> PERSISTENT_REGISTERS = Common::X64::BuildRegSet({
     BRANCH_HOLDER,
 });
 
-MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d_)
-    : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {}
+// Arbitrarily chosen based on current booting games.
+constexpr size_t MAX_CODE_SIZE = 0x10000;
 
-std::unique_ptr<CachedMacro> MacroJITx64::Compile(const std::vector<u32>& code) {
-    return std::make_unique<MacroJITx64Impl>(maxwell3d, code);
+std::bitset<32> PersistentCallerSavedRegs() {
+    return PERSISTENT_REGISTERS & Common::X64::ABI_ALL_CALLER_SAVED;
 }
 
-MacroJITx64Impl::MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_)
-    : CodeGenerator{MAX_CODE_SIZE}, code{code_}, maxwell3d{maxwell3d_} {
-    Compile();
-}
+class MacroJITx64Impl final : public Xbyak::CodeGenerator, public CachedMacro {
+public:
+    explicit MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_)
+        : CodeGenerator{MAX_CODE_SIZE}, code{code_}, maxwell3d{maxwell3d_} {
+        Compile();
+    }
+
+    void Execute(const std::vector<u32>& parameters, u32 method) override;
+
+    void Compile_ALU(Macro::Opcode opcode);
+    void Compile_AddImmediate(Macro::Opcode opcode);
+    void Compile_ExtractInsert(Macro::Opcode opcode);
+    void Compile_ExtractShiftLeftImmediate(Macro::Opcode opcode);
+    void Compile_ExtractShiftLeftRegister(Macro::Opcode opcode);
+    void Compile_Read(Macro::Opcode opcode);
+    void Compile_Branch(Macro::Opcode opcode);
+
+private:
+    void Optimizer_ScanFlags();
+
+    void Compile();
+    bool Compile_NextInstruction();
+
+    Xbyak::Reg32 Compile_FetchParameter();
+    Xbyak::Reg32 Compile_GetRegister(u32 index, Xbyak::Reg32 dst);
+
+    void Compile_ProcessResult(Macro::ResultOperation operation, u32 reg);
+    void Compile_Send(Xbyak::Reg32 value);
 
-MacroJITx64Impl::~MacroJITx64Impl() = default;
+    Macro::Opcode GetOpCode() const;
+
+    struct JITState {
+        Engines::Maxwell3D* maxwell3d{};
+        std::array<u32, Macro::NUM_MACRO_REGISTERS> registers{};
+        u32 carry_flag{};
+    };
+    static_assert(offsetof(JITState, maxwell3d) == 0, "Maxwell3D is not at 0x0");
+    using ProgramType = void (*)(JITState*, const u32*);
+
+    struct OptimizerState {
+        bool can_skip_carry{};
+        bool has_delayed_pc{};
+        bool zero_reg_skip{};
+        bool skip_dummy_addimmediate{};
+        bool optimize_for_method_move{};
+        bool enable_asserts{};
+    };
+    OptimizerState optimizer{};
+
+    std::optional<Macro::Opcode> next_opcode{};
+    ProgramType program{nullptr};
+
+    std::array<Xbyak::Label, MAX_CODE_SIZE> labels;
+    std::array<Xbyak::Label, MAX_CODE_SIZE> delay_skip;
+    Xbyak::Label end_of_code{};
+
+    bool is_delay_slot{};
+    u32 pc{};
+
+    const std::vector<u32>& code;
+    Engines::Maxwell3D& maxwell3d;
+};
 
 void MacroJITx64Impl::Execute(const std::vector<u32>& parameters, u32 method) {
     MICROPROFILE_SCOPE(MacroJitExecute);
@@ -307,11 +372,11 @@ void MacroJITx64Impl::Compile_Read(Macro::Opcode opcode) {
     Compile_ProcessResult(opcode.result_operation, opcode.dst);
 }
 
-static void Send(Engines::Maxwell3D* maxwell3d, Macro::MethodAddress method_address, u32 value) {
+void Send(Engines::Maxwell3D* maxwell3d, Macro::MethodAddress method_address, u32 value) {
     maxwell3d->CallMethodFromMME(method_address.address, value);
 }
 
-void Tegra::MacroJITx64Impl::Compile_Send(Xbyak::Reg32 value) {
+void MacroJITx64Impl::Compile_Send(Xbyak::Reg32 value) {
     Common::X64::ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
     mov(Common::X64::ABI_PARAM1, qword[STATE]);
     mov(Common::X64::ABI_PARAM2, METHOD_ADDRESS);
@@ -338,7 +403,7 @@ void Tegra::MacroJITx64Impl::Compile_Send(Xbyak::Reg32 value) {
     L(dont_process);
 }
 
-void Tegra::MacroJITx64Impl::Compile_Branch(Macro::Opcode opcode) {
+void MacroJITx64Impl::Compile_Branch(Macro::Opcode opcode) {
     ASSERT_MSG(!is_delay_slot, "Executing a branch in a delay slot is not valid");
     const s32 jump_address =
         static_cast<s32>(pc) + static_cast<s32>(opcode.GetBranchTarget() / sizeof(s32));
@@ -392,7 +457,7 @@ void Tegra::MacroJITx64Impl::Compile_Branch(Macro::Opcode opcode) {
     L(end);
 }
 
-void Tegra::MacroJITx64Impl::Optimizer_ScanFlags() {
+void MacroJITx64Impl::Optimizer_ScanFlags() {
     optimizer.can_skip_carry = true;
     optimizer.has_delayed_pc = false;
     for (auto raw_op : code) {
@@ -534,7 +599,7 @@ bool MacroJITx64Impl::Compile_NextInstruction() {
     return true;
 }
 
-Xbyak::Reg32 Tegra::MacroJITx64Impl::Compile_FetchParameter() {
+Xbyak::Reg32 MacroJITx64Impl::Compile_FetchParameter() {
     mov(eax, dword[PARAMETERS]);
     add(PARAMETERS, sizeof(u32));
     return eax;
@@ -611,9 +676,12 @@ Macro::Opcode MacroJITx64Impl::GetOpCode() const {
     ASSERT(pc < code.size());
     return {code[pc]};
 }
+} // Anonymous namespace
 
-std::bitset<32> MacroJITx64Impl::PersistentCallerSavedRegs() const {
-    return PERSISTENT_REGISTERS & Common::X64::ABI_ALL_CALLER_SAVED;
-}
+MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d_)
+    : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {}
 
+std::unique_ptr<CachedMacro> MacroJITx64::Compile(const std::vector<u32>& code) {
+    return std::make_unique<MacroJITx64Impl>(maxwell3d, code);
+}
 } // namespace Tegra
diff --git a/src/video_core/macro/macro_jit_x64.h b/src/video_core/macro/macro_jit_x64.h
index d03d480b4..773b037ae 100644
--- a/src/video_core/macro/macro_jit_x64.h
+++ b/src/video_core/macro/macro_jit_x64.h
@@ -4,12 +4,7 @@
 
 #pragma once
 
-#include <array>
-#include <bitset>
-#include <xbyak/xbyak.h>
-#include "common/bit_field.h"
 #include "common/common_types.h"
-#include "common/x64/xbyak_abi.h"
 #include "video_core/macro/macro.h"
 
 namespace Tegra {
@@ -18,9 +13,6 @@ namespace Engines {
 class Maxwell3D;
 }
 
-/// MAX_CODE_SIZE is arbitrarily chosen based on current booting games
-constexpr size_t MAX_CODE_SIZE = 0x10000;
-
 class MacroJITx64 final : public MacroEngine {
 public:
     explicit MacroJITx64(Engines::Maxwell3D& maxwell3d_);
@@ -32,67 +24,4 @@ private:
     Engines::Maxwell3D& maxwell3d;
 };
 
-class MacroJITx64Impl : public Xbyak::CodeGenerator, public CachedMacro {
-public:
-    explicit MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_);
-    ~MacroJITx64Impl();
-
-    void Execute(const std::vector<u32>& parameters, u32 method) override;
-
-    void Compile_ALU(Macro::Opcode opcode);
-    void Compile_AddImmediate(Macro::Opcode opcode);
-    void Compile_ExtractInsert(Macro::Opcode opcode);
-    void Compile_ExtractShiftLeftImmediate(Macro::Opcode opcode);
-    void Compile_ExtractShiftLeftRegister(Macro::Opcode opcode);
-    void Compile_Read(Macro::Opcode opcode);
-    void Compile_Branch(Macro::Opcode opcode);
-
-private:
-    void Optimizer_ScanFlags();
-
-    void Compile();
-    bool Compile_NextInstruction();
-
-    Xbyak::Reg32 Compile_FetchParameter();
-    Xbyak::Reg32 Compile_GetRegister(u32 index, Xbyak::Reg32 dst);
-
-    void Compile_ProcessResult(Macro::ResultOperation operation, u32 reg);
-    void Compile_Send(Xbyak::Reg32 value);
-
-    Macro::Opcode GetOpCode() const;
-    std::bitset<32> PersistentCallerSavedRegs() const;
-
-    struct JITState {
-        Engines::Maxwell3D* maxwell3d{};
-        std::array<u32, Macro::NUM_MACRO_REGISTERS> registers{};
-        u32 carry_flag{};
-    };
-    static_assert(offsetof(JITState, maxwell3d) == 0, "Maxwell3D is not at 0x0");
-    using ProgramType = void (*)(JITState*, const u32*);
-
-    struct OptimizerState {
-        bool can_skip_carry{};
-        bool has_delayed_pc{};
-        bool zero_reg_skip{};
-        bool skip_dummy_addimmediate{};
-        bool optimize_for_method_move{};
-        bool enable_asserts{};
-    };
-    OptimizerState optimizer{};
-
-    std::optional<Macro::Opcode> next_opcode{};
-    ProgramType program{nullptr};
-
-    std::array<Xbyak::Label, MAX_CODE_SIZE> labels;
-    std::array<Xbyak::Label, MAX_CODE_SIZE> delay_skip;
-    Xbyak::Label end_of_code{};
-
-    bool is_delay_slot{};
-    u32 pc{};
-    std::optional<u32> delayed_pc;
-
-    const std::vector<u32>& code;
-    Engines::Maxwell3D& maxwell3d;
-};
-
 } // namespace Tegra
diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h
index 38d8d9d74..61bfe47c7 100644
--- a/src/video_core/memory_manager.h
+++ b/src/video_core/memory_manager.h
@@ -143,6 +143,8 @@ public:
     [[nodiscard]] GPUVAddr Allocate(std::size_t size, std::size_t align);
     void Unmap(GPUVAddr gpu_addr, std::size_t size);
 
+    void FlushRegion(GPUVAddr gpu_addr, size_t size) const;
+
 private:
     [[nodiscard]] PageEntry GetPageEntry(GPUVAddr gpu_addr) const;
     void SetPageEntry(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size = page_size);
@@ -153,8 +155,6 @@ private:
     void TryLockPage(PageEntry page_entry, std::size_t size);
     void TryUnlockPage(PageEntry page_entry, std::size_t size);
 
-    void FlushRegion(GPUVAddr gpu_addr, size_t size) const;
-
     void ReadBlockImpl(GPUVAddr gpu_src_addr, void* dest_buffer, std::size_t size,
                        bool is_safe) const;
     void WriteBlockImpl(GPUVAddr gpu_dest_addr, const void* src_buffer, std::size_t size,
diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h
index b094fc064..1f1f12291 100644
--- a/src/video_core/rasterizer_interface.h
+++ b/src/video_core/rasterizer_interface.h
@@ -123,6 +123,9 @@ public:
 
     [[nodiscard]] virtual Tegra::Engines::AccelerateDMAInterface& AccessAccelerateDMA() = 0;
 
+    virtual void AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                          std::span<u8> memory) = 0;
+
     /// Attempt to use a faster method to display the framebuffer to screen
     [[nodiscard]] virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config,
                                                  VAddr framebuffer_addr, u32 pixel_stride) {
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 9b516c64f..142412a8e 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -484,6 +484,28 @@ Tegra::Engines::AccelerateDMAInterface& RasterizerOpenGL::AccessAccelerateDMA()
     return accelerate_dma;
 }
 
+void RasterizerOpenGL::AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                                std::span<u8> memory) {
+    auto cpu_addr = gpu_memory.GpuToCpuAddress(address);
+    if (!cpu_addr) [[unlikely]] {
+        gpu_memory.WriteBlock(address, memory.data(), copy_size);
+        return;
+    }
+    gpu_memory.WriteBlockUnsafe(address, memory.data(), copy_size);
+    {
+        std::unique_lock<std::mutex> lock{buffer_cache.mutex};
+        if (!buffer_cache.InlineMemory(*cpu_addr, copy_size, memory)) {
+            buffer_cache.WriteMemory(*cpu_addr, copy_size);
+        }
+    }
+    {
+        std::scoped_lock lock_texture{texture_cache.mutex};
+        texture_cache.WriteMemory(*cpu_addr, copy_size);
+    }
+    shader_cache.InvalidateRegion(*cpu_addr, copy_size);
+    query_cache.InvalidateRegion(*cpu_addr, copy_size);
+}
+
 bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
                                          VAddr framebuffer_addr, u32 pixel_stride) {
     if (framebuffer_addr == 0) {
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index d0397b745..98f6fd342 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -106,6 +106,8 @@ public:
                                const Tegra::Engines::Fermi2D::Surface& dst,
                                const Tegra::Engines::Fermi2D::Config& copy_config) override;
     Tegra::Engines::AccelerateDMAInterface& AccessAccelerateDMA() override;
+    void AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                  std::span<u8> memory) override;
     bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
                            u32 pixel_stride) override;
     void LoadDiskResources(u64 title_id, std::stop_token stop_loading,
diff --git a/src/video_core/renderer_vulkan/vk_fsr.cpp b/src/video_core/renderer_vulkan/vk_fsr.cpp
index 73629d229..b630090e8 100644
--- a/src/video_core/renderer_vulkan/vk_fsr.cpp
+++ b/src/video_core/renderer_vulkan/vk_fsr.cpp
@@ -214,7 +214,7 @@ VkImageView FSR::Draw(VKScheduler& scheduler, size_t image_index, VkImageView im
 
         {
             VkImageMemoryBarrier fsr_write_barrier = base_barrier;
-            fsr_write_barrier.image = *images[image_index],
+            fsr_write_barrier.image = *images[image_index];
             fsr_write_barrier.oldLayout = VK_IMAGE_LAYOUT_UNDEFINED;
 
             cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT,
diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp
index fd334a146..2227d9197 100644
--- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp
+++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp
@@ -548,6 +548,28 @@ Tegra::Engines::AccelerateDMAInterface& RasterizerVulkan::AccessAccelerateDMA()
     return accelerate_dma;
 }
 
+void RasterizerVulkan::AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                                std::span<u8> memory) {
+    auto cpu_addr = gpu_memory.GpuToCpuAddress(address);
+    if (!cpu_addr) [[unlikely]] {
+        gpu_memory.WriteBlock(address, memory.data(), copy_size);
+        return;
+    }
+    gpu_memory.WriteBlockUnsafe(address, memory.data(), copy_size);
+    {
+        std::unique_lock<std::mutex> lock{buffer_cache.mutex};
+        if (!buffer_cache.InlineMemory(*cpu_addr, copy_size, memory)) {
+            buffer_cache.WriteMemory(*cpu_addr, copy_size);
+        }
+    }
+    {
+        std::scoped_lock lock_texture{texture_cache.mutex};
+        texture_cache.WriteMemory(*cpu_addr, copy_size);
+    }
+    pipeline_cache.InvalidateRegion(*cpu_addr, copy_size);
+    query_cache.InvalidateRegion(*cpu_addr, copy_size);
+}
+
 bool RasterizerVulkan::AccelerateDisplay(const Tegra::FramebufferConfig& config,
                                          VAddr framebuffer_addr, u32 pixel_stride) {
     if (!framebuffer_addr) {
diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h
index 866827247..5af2e275b 100644
--- a/src/video_core/renderer_vulkan/vk_rasterizer.h
+++ b/src/video_core/renderer_vulkan/vk_rasterizer.h
@@ -99,6 +99,8 @@ public:
                                const Tegra::Engines::Fermi2D::Surface& dst,
                                const Tegra::Engines::Fermi2D::Config& copy_config) override;
     Tegra::Engines::AccelerateDMAInterface& AccessAccelerateDMA() override;
+    void AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                  std::span<u8> memory) override;
     bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
                            u32 pixel_stride) override;
     void LoadDiskResources(u64 title_id, std::stop_token stop_loading,
diff --git a/src/video_core/shader_cache.cpp b/src/video_core/shader_cache.cpp
index 78bf90c48..87636857d 100644
--- a/src/video_core/shader_cache.cpp
+++ b/src/video_core/shader_cache.cpp
@@ -170,7 +170,7 @@ void ShaderCache::RemovePendingShaders() {
     marked_for_removal.clear();
 
     if (!removed_shaders.empty()) {
-        RemoveShadersFromStorage(std::move(removed_shaders));
+        RemoveShadersFromStorage(removed_shaders);
     }
 }
 
@@ -213,7 +213,7 @@ void ShaderCache::UnmarkMemory(Entry* entry) {
     rasterizer.UpdatePagesCachedCount(addr, size, -1);
 }
 
-void ShaderCache::RemoveShadersFromStorage(std::vector<ShaderInfo*> removed_shaders) {
+void ShaderCache::RemoveShadersFromStorage(std::span<ShaderInfo*> removed_shaders) {
     // Remove them from the cache
     std::erase_if(storage, [&removed_shaders](const std::unique_ptr<ShaderInfo>& shader) {
         return std::ranges::find(removed_shaders, shader.get()) != removed_shaders.end();
diff --git a/src/video_core/shader_cache.h b/src/video_core/shader_cache.h
index 136fe294c..8836bc8c6 100644
--- a/src/video_core/shader_cache.h
+++ b/src/video_core/shader_cache.h
@@ -4,7 +4,6 @@
 
 #pragma once
 
-#include <algorithm>
 #include <array>
 #include <memory>
 #include <mutex>
@@ -138,7 +137,7 @@ private:
     /// @param removed_shaders Shaders to be removed from the storage
     /// @pre invalidation_mutex is locked
     /// @pre lookup_mutex is locked
-    void RemoveShadersFromStorage(std::vector<ShaderInfo*> removed_shaders);
+    void RemoveShadersFromStorage(std::span<ShaderInfo*> removed_shaders);
 
     /// @brief Creates a new entry in the lookup cache and returns its pointer
     /// @pre lookup_mutex is locked
diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp
index 7e8f92840..2c8c10c50 100644
--- a/src/yuzu/configuration/config.cpp
+++ b/src/yuzu/configuration/config.cpp
@@ -393,6 +393,8 @@ void Config::ReadControlValues() {
     ReadGlobalSetting(Settings::values.enable_accurate_vibrations);
     ReadGlobalSetting(Settings::values.motion_enabled);
 
+    ReadBasicSetting(Settings::values.controller_navigation);
+
     qt_config->endGroup();
 }
 
@@ -1001,6 +1003,7 @@ void Config::SaveControlValues() {
     WriteBasicSetting(Settings::values.keyboard_enabled);
     WriteBasicSetting(Settings::values.emulate_analog_keyboard);
     WriteBasicSetting(Settings::values.mouse_panning_sensitivity);
+    WriteBasicSetting(Settings::values.controller_navigation);
 
     WriteBasicSetting(Settings::values.tas_enable);
     WriteBasicSetting(Settings::values.tas_loop);
diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui
index 9241678e4..74f0e0b79 100644
--- a/src/yuzu/configuration/configure_graphics.ui
+++ b/src/yuzu/configuration/configure_graphics.ui
@@ -429,7 +429,7 @@
              </item>
              <item>
               <property name="text">
-               <string>AMD FidelityFX™️ Super Resolution [Vulkan Only]</string>
+               <string>AMD FidelityFX™️ Super Resolution (Vulkan Only)</string>
               </property>
              </item>
             </widget>
diff --git a/src/yuzu/configuration/configure_input_advanced.cpp b/src/yuzu/configuration/configure_input_advanced.cpp
index 65c8e59ac..20fc2599d 100644
--- a/src/yuzu/configuration/configure_input_advanced.cpp
+++ b/src/yuzu/configuration/configure_input_advanced.cpp
@@ -131,6 +131,7 @@ void ConfigureInputAdvanced::ApplyConfiguration() {
     Settings::values.touchscreen.enabled = ui->touchscreen_enabled->isChecked();
     Settings::values.enable_raw_input = ui->enable_raw_input->isChecked();
     Settings::values.enable_udp_controller = ui->enable_udp_controller->isChecked();
+    Settings::values.controller_navigation = ui->controller_navigation->isChecked();
 }
 
 void ConfigureInputAdvanced::LoadConfiguration() {
@@ -162,6 +163,7 @@ void ConfigureInputAdvanced::LoadConfiguration() {
     ui->touchscreen_enabled->setChecked(Settings::values.touchscreen.enabled);
     ui->enable_raw_input->setChecked(Settings::values.enable_raw_input.GetValue());
     ui->enable_udp_controller->setChecked(Settings::values.enable_udp_controller.GetValue());
+    ui->controller_navigation->setChecked(Settings::values.controller_navigation.GetValue());
 
     UpdateUIEnabled();
 }
diff --git a/src/yuzu/configuration/configure_input_advanced.ui b/src/yuzu/configuration/configure_input_advanced.ui
index df0e4d602..66f2075f2 100644
--- a/src/yuzu/configuration/configure_input_advanced.ui
+++ b/src/yuzu/configuration/configure_input_advanced.ui
@@ -2655,6 +2655,19 @@
                      </widget>
                    </item>
                    <item row="4" column="0">
+                     <widget class="QCheckBox" name="controller_navigation">
+                       <property name="minimumSize">
+                         <size>
+                           <width>0</width>
+                           <height>23</height>
+                         </size>
+                       </property>
+                       <property name="text">
+                         <string>Controller navigation</string>
+                       </property>
+                     </widget>
+                   </item>
+                   <item row="5" column="0">
                      <widget class="QCheckBox" name="mouse_panning">
                        <property name="minimumSize">
                          <size>
@@ -2667,7 +2680,7 @@
                        </property>
                      </widget>
                    </item>
-                   <item row="4" column="2">
+                   <item row="5" column="2">
                      <widget class="QSpinBox" name="mouse_panning_sensitivity">
                        <property name="toolTip">
                          <string>Mouse sensitivity</string>
@@ -2689,14 +2702,14 @@
                        </property>
                      </widget>
                    </item>
-                   <item row="5" column="0">
+                   <item row="6" column="0">
                      <widget class="QLabel" name="motion_touch">
                        <property name="text">
                          <string>Motion / Touch</string>
                        </property>
                      </widget>
                    </item>
-                   <item row="5" column="2">
+                   <item row="6" column="2">
                      <widget class="QPushButton" name="buttonMotionTouch">
                        <property name="text">
                          <string>Configure</string>
diff --git a/src/yuzu/configuration/configure_input_player.cpp b/src/yuzu/configuration/configure_input_player.cpp
index d2132b408..ba563b382 100644
--- a/src/yuzu/configuration/configure_input_player.cpp
+++ b/src/yuzu/configuration/configure_input_player.cpp
@@ -147,7 +147,7 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) {
     // Retrieve the names from Qt
     if (param.Get("engine", "") == "keyboard") {
         const QString button_str = GetKeyName(param.Get("code", 0));
-        return QObject::tr("%1%2").arg(toggle, button_str);
+        return QObject::tr("%1%2%3").arg(toggle, inverted, button_str);
     }
 
     if (common_button_name == Common::Input::ButtonNames::Invalid) {
@@ -326,7 +326,7 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
         connect(button, &QPushButton::clicked, [=, this] {
             HandleClick(
                 button, button_id,
-                [=, this](Common::ParamPackage params) {
+                [=, this](const Common::ParamPackage& params) {
                     emulated_controller->SetButtonParam(button_id, params);
                 },
                 InputCommon::Polling::InputType::Button);
@@ -341,7 +341,7 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
                         emulated_controller->SetButtonParam(button_id, {});
                         button_map[button_id]->setText(tr("[not set]"));
                     });
-                    if (param.Has("button") || param.Has("hat")) {
+                    if (param.Has("code") || param.Has("button") || param.Has("hat")) {
                         context_menu.addAction(tr("Toggle button"), [&] {
                             const bool toggle_value = !param.Get("toggle", false);
                             param.Set("toggle", toggle_value);
@@ -349,8 +349,8 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
                             emulated_controller->SetButtonParam(button_id, param);
                         });
                         context_menu.addAction(tr("Invert button"), [&] {
-                            const bool toggle_value = !param.Get("inverted", false);
-                            param.Set("inverted", toggle_value);
+                            const bool invert_value = !param.Get("inverted", false);
+                            param.Set("inverted", invert_value);
                             button_map[button_id]->setText(ButtonToText(param));
                             emulated_controller->SetButtonParam(button_id, param);
                         });
@@ -392,7 +392,7 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
         connect(button, &QPushButton::clicked, [=, this] {
             HandleClick(
                 button, motion_id,
-                [=, this](Common::ParamPackage params) {
+                [=, this](const Common::ParamPackage& params) {
                     emulated_controller->SetMotionParam(motion_id, params);
                 },
                 InputCommon::Polling::InputType::Motion);
@@ -403,10 +403,22 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
         connect(button, &QPushButton::customContextMenuRequested,
                 [=, this](const QPoint& menu_location) {
                     QMenu context_menu;
+                    Common::ParamPackage param = emulated_controller->GetMotionParam(motion_id);
                     context_menu.addAction(tr("Clear"), [&] {
                         emulated_controller->SetMotionParam(motion_id, {});
                         motion_map[motion_id]->setText(tr("[not set]"));
                     });
+                    if (param.Has("motion")) {
+                        context_menu.addAction(tr("Set gyro threshold"), [&] {
+                            const int gyro_threshold =
+                                static_cast<int>(param.Get("threshold", 0.007f) * 1000.0f);
+                            const int new_threshold = QInputDialog::getInt(
+                                this, tr("Set threshold"), tr("Choose a value between 0% and 100%"),
+                                gyro_threshold, 0, 100);
+                            param.Set("threshold", new_threshold / 1000.0f);
+                            emulated_controller->SetMotionParam(motion_id, param);
+                        });
+                    }
                     context_menu.exec(motion_map[motion_id]->mapToGlobal(menu_location));
                 });
     }
@@ -485,10 +497,11 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
                                 param.Set("invert_y", invert_str);
                                 emulated_controller->SetStickParam(analog_id, param);
                             }
-                            for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM;
-                                 ++sub_button_id) {
-                                analog_map_buttons[analog_id][sub_button_id]->setText(
-                                    AnalogToText(param, analog_sub_buttons[sub_button_id]));
+                            for (int analog_sub_button_id = 0;
+                                 analog_sub_button_id < ANALOG_SUB_BUTTONS_NUM;
+                                 ++analog_sub_button_id) {
+                                analog_map_buttons[analog_id][analog_sub_button_id]->setText(
+                                    AnalogToText(param, analog_sub_buttons[analog_sub_button_id]));
                             }
                         });
                         context_menu.exec(analog_map_buttons[analog_id][sub_button_id]->mapToGlobal(
@@ -510,28 +523,37 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
 
         analog_map_modifier_button[analog_id]->setContextMenuPolicy(Qt::CustomContextMenu);
 
-        connect(analog_map_modifier_button[analog_id], &QPushButton::customContextMenuRequested,
-                [=, this](const QPoint& menu_location) {
-                    QMenu context_menu;
-                    Common::ParamPackage param = emulated_controller->GetStickParam(analog_id);
-                    context_menu.addAction(tr("Clear"), [&] {
-                        param.Set("modifier", "");
-                        analog_map_modifier_button[analog_id]->setText(tr("[not set]"));
-                        emulated_controller->SetStickParam(analog_id, param);
-                    });
-                    context_menu.addAction(tr("Toggle button"), [&] {
-                        Common::ParamPackage modifier_param =
-                            Common::ParamPackage{param.Get("modifier", "")};
-                        const bool toggle_value = !modifier_param.Get("toggle", false);
-                        modifier_param.Set("toggle", toggle_value);
-                        param.Set("modifier", modifier_param.Serialize());
-                        analog_map_modifier_button[analog_id]->setText(
-                            ButtonToText(modifier_param));
-                        emulated_controller->SetStickParam(analog_id, param);
-                    });
-                    context_menu.exec(
-                        analog_map_modifier_button[analog_id]->mapToGlobal(menu_location));
+        connect(
+            analog_map_modifier_button[analog_id], &QPushButton::customContextMenuRequested,
+            [=, this](const QPoint& menu_location) {
+                QMenu context_menu;
+                Common::ParamPackage param = emulated_controller->GetStickParam(analog_id);
+                context_menu.addAction(tr("Clear"), [&] {
+                    param.Set("modifier", "");
+                    analog_map_modifier_button[analog_id]->setText(tr("[not set]"));
+                    emulated_controller->SetStickParam(analog_id, param);
                 });
+                context_menu.addAction(tr("Toggle button"), [&] {
+                    Common::ParamPackage modifier_param =
+                        Common::ParamPackage{param.Get("modifier", "")};
+                    const bool toggle_value = !modifier_param.Get("toggle", false);
+                    modifier_param.Set("toggle", toggle_value);
+                    param.Set("modifier", modifier_param.Serialize());
+                    analog_map_modifier_button[analog_id]->setText(ButtonToText(modifier_param));
+                    emulated_controller->SetStickParam(analog_id, param);
+                });
+                context_menu.addAction(tr("Invert button"), [&] {
+                    Common::ParamPackage modifier_param =
+                        Common::ParamPackage{param.Get("modifier", "")};
+                    const bool invert_value = !modifier_param.Get("inverted", false);
+                    modifier_param.Set("inverted", invert_value);
+                    param.Set("modifier", modifier_param.Serialize());
+                    analog_map_modifier_button[analog_id]->setText(ButtonToText(modifier_param));
+                    emulated_controller->SetStickParam(analog_id, param);
+                });
+                context_menu.exec(
+                    analog_map_modifier_button[analog_id]->mapToGlobal(menu_location));
+            });
 
         connect(analog_map_range_spinbox[analog_id], qOverload<int>(&QSpinBox::valueChanged),
                 [=, this] {
@@ -762,7 +784,7 @@ void ConfigureInputPlayer::UpdateInputDeviceCombobox() {
     if (devices.size() == 1) {
         const auto devices_it = std::find_if(
             input_devices.begin(), input_devices.end(),
-            [first_engine, first_guid, first_port, first_pad](const Common::ParamPackage param) {
+            [first_engine, first_guid, first_port, first_pad](const Common::ParamPackage& param) {
                 return param.Get("engine", "") == first_engine &&
                        param.Get("guid", "") == first_guid && param.Get("port", 0) == first_port &&
                        param.Get("pad", 0) == first_pad;
@@ -793,7 +815,7 @@ void ConfigureInputPlayer::UpdateInputDeviceCombobox() {
     if (is_engine_equal && is_port_equal) {
         const auto devices_it = std::find_if(
             input_devices.begin(), input_devices.end(),
-            [first_engine, first_guid, second_guid, first_port](const Common::ParamPackage param) {
+            [first_engine, first_guid, second_guid, first_port](const Common::ParamPackage& param) {
                 const bool is_guid_valid =
                     (param.Get("guid", "") == first_guid &&
                      param.Get("guid2", "") == second_guid) ||
@@ -1005,7 +1027,7 @@ int ConfigureInputPlayer::GetIndexFromControllerType(Core::HID::NpadStyleIndex t
 void ConfigureInputPlayer::UpdateInputDevices() {
     input_devices = input_subsystem->GetInputDevices();
     ui->comboDevices->clear();
-    for (auto device : input_devices) {
+    for (const auto& device : input_devices) {
         ui->comboDevices->addItem(QString::fromStdString(device.Get("display", "Unknown")), {});
     }
 }
@@ -1287,7 +1309,7 @@ void ConfigureInputPlayer::HandleClick(
     }
     button->setFocus();
 
-    input_setter = new_input_setter;
+    input_setter = std::move(new_input_setter);
 
     input_subsystem->BeginMapping(type);
 
@@ -1337,7 +1359,7 @@ bool ConfigureInputPlayer::IsInputAcceptable(const Common::ParamPackage& params)
         return params.Get("engine", "") == "keyboard" || params.Get("engine", "") == "mouse";
     }
 
-    const auto current_input_device = input_devices[ui->comboDevices->currentIndex()];
+    const auto& current_input_device = input_devices[ui->comboDevices->currentIndex()];
     return params.Get("engine", "") == current_input_device.Get("engine", "") &&
            (params.Get("guid", "") == current_input_device.Get("guid", "") ||
             params.Get("guid", "") == current_input_device.Get("guid2", "")) &&
diff --git a/src/yuzu/configuration/configure_motion_touch.cpp b/src/yuzu/configuration/configure_motion_touch.cpp
index 8539a5c8b..4340de304 100644
--- a/src/yuzu/configuration/configure_motion_touch.cpp
+++ b/src/yuzu/configuration/configure_motion_touch.cpp
@@ -42,23 +42,25 @@ CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent,
     job = std::make_unique<CalibrationConfigurationJob>(
         host, port,
         [this](CalibrationConfigurationJob::Status status) {
-            QString text;
-            switch (status) {
-            case CalibrationConfigurationJob::Status::Ready:
-                text = tr("Touch the top left corner <br>of your touchpad.");
-                break;
-            case CalibrationConfigurationJob::Status::Stage1Completed:
-                text = tr("Now touch the bottom right corner <br>of your touchpad.");
-                break;
-            case CalibrationConfigurationJob::Status::Completed:
-                text = tr("Configuration completed!");
-                break;
-            default:
-                break;
-            }
-            QMetaObject::invokeMethod(this, "UpdateLabelText", Q_ARG(QString, text));
+            QMetaObject::invokeMethod(this, [status, this] {
+                QString text;
+                switch (status) {
+                case CalibrationConfigurationJob::Status::Ready:
+                    text = tr("Touch the top left corner <br>of your touchpad.");
+                    break;
+                case CalibrationConfigurationJob::Status::Stage1Completed:
+                    text = tr("Now touch the bottom right corner <br>of your touchpad.");
+                    break;
+                case CalibrationConfigurationJob::Status::Completed:
+                    text = tr("Configuration completed!");
+                    break;
+                default:
+                    break;
+                }
+                UpdateLabelText(text);
+            });
             if (status == CalibrationConfigurationJob::Status::Completed) {
-                QMetaObject::invokeMethod(this, "UpdateButtonText", Q_ARG(QString, tr("OK")));
+                QMetaObject::invokeMethod(this, [this] { UpdateButtonText(tr("OK")); });
             }
         },
         [this](u16 min_x_, u16 min_y_, u16 max_x_, u16 max_y_) {
@@ -215,11 +217,11 @@ void ConfigureMotionTouch::OnCemuhookUDPTest() {
         ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()),
         [this] {
             LOG_INFO(Frontend, "UDP input test success");
-            QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, true));
+            QMetaObject::invokeMethod(this, [this] { ShowUDPTestResult(true); });
         },
         [this] {
             LOG_ERROR(Frontend, "UDP input test failed");
-            QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, false));
+            QMetaObject::invokeMethod(this, [this] { ShowUDPTestResult(false); });
         });
 }
 
diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp
index 8b5c4a10a..e3661b390 100644
--- a/src/yuzu/game_list.cpp
+++ b/src/yuzu/game_list.cpp
@@ -161,7 +161,7 @@ GameListSearchField::GameListSearchField(GameList* parent) : QWidget{parent} {
  * @return true if the haystack contains all words of userinput
  */
 static bool ContainsAllWords(const QString& haystack, const QString& userinput) {
-    const QStringList userinput_split = userinput.split(QLatin1Char{' '}, QString::SkipEmptyParts);
+    const QStringList userinput_split = userinput.split(QLatin1Char{' '}, Qt::SkipEmptyParts);
 
     return std::all_of(userinput_split.begin(), userinput_split.end(),
                        [&haystack](const QString& s) { return haystack.contains(s); });
diff --git a/src/yuzu/hotkeys.cpp b/src/yuzu/hotkeys.cpp
index d96497c4e..6ed9611c7 100644
--- a/src/yuzu/hotkeys.cpp
+++ b/src/yuzu/hotkeys.cpp
@@ -190,6 +190,9 @@ void ControllerShortcut::ControllerUpdateEvent(Core::HID::ControllerTriggerType
     if (type != Core::HID::ControllerTriggerType::Button) {
         return;
     }
+    if (!Settings::values.controller_navigation) {
+        return;
+    }
     if (button_sequence.npad.raw == Core::HID::NpadButton::None &&
         button_sequence.capture.raw == 0 && button_sequence.home.raw == 0) {
         return;
diff --git a/src/yuzu/util/controller_navigation.cpp b/src/yuzu/util/controller_navigation.cpp
index 86fb28b9f..c2b13123d 100644
--- a/src/yuzu/util/controller_navigation.cpp
+++ b/src/yuzu/util/controller_navigation.cpp
@@ -40,6 +40,9 @@ void ControllerNavigation::TriggerButton(Settings::NativeButton::Values native_b
 
 void ControllerNavigation::ControllerUpdateEvent(Core::HID::ControllerTriggerType type) {
     std::lock_guard lock{mutex};
+    if (!Settings::values.controller_navigation) {
+        return;
+    }
     if (type == Core::HID::ControllerTriggerType::Button) {
         ControllerUpdateButton();
         return;